Merge branch 'main' into tmp-release-v1.11.0

crates/meilisearch/src/analytics/mock_analytics.rs

@@ -0,0 +1,109 @@
+use std::any::Any;
+use std::sync::Arc;
+
+use actix_web::HttpRequest;
+use meilisearch_types::InstanceUid;
+use serde_json::Value;
+
+use super::{find_user_id, Analytics, DocumentDeletionKind, DocumentFetchKind};
+use crate::routes::indexes::documents::{DocumentEditionByFunction, UpdateDocumentsQuery};
+use crate::Opt;
+
+pub struct MockAnalytics {
+    instance_uid: Option<InstanceUid>,
+}
+
+#[derive(Default)]
+pub struct SearchAggregator;
+
+#[allow(dead_code)]
+impl SearchAggregator {
+    pub fn from_query(_: &dyn Any, _: &dyn Any) -> Self {
+        Self
+    }
+
+    pub fn succeed(&mut self, _: &dyn Any) {}
+}
+
+#[derive(Default)]
+pub struct SimilarAggregator;
+
+#[allow(dead_code)]
+impl SimilarAggregator {
+    pub fn from_query(_: &dyn Any, _: &dyn Any) -> Self {
+        Self
+    }
+
+    pub fn succeed(&mut self, _: &dyn Any) {}
+}
+
+#[derive(Default)]
+pub struct MultiSearchAggregator;
+
+#[allow(dead_code)]
+impl MultiSearchAggregator {
+    pub fn from_federated_search(_: &dyn Any, _: &dyn Any) -> Self {
+        Self
+    }
+
+    pub fn succeed(&mut self) {}
+}
+
+#[derive(Default)]
+pub struct FacetSearchAggregator;
+
+#[allow(dead_code)]
+impl FacetSearchAggregator {
+    pub fn from_query(_: &dyn Any, _: &dyn Any) -> Self {
+        Self
+    }
+
+    pub fn succeed(&mut self, _: &dyn Any) {}
+}
+
+impl MockAnalytics {
+    #[allow(clippy::new_ret_no_self)]
+    pub fn new(opt: &Opt) -> Arc<dyn Analytics> {
+        let instance_uid = find_user_id(&opt.db_path);
+        Arc::new(Self { instance_uid })
+    }
+}
+
+impl Analytics for MockAnalytics {
+    fn instance_uid(&self) -> Option<&meilisearch_types::InstanceUid> {
+        self.instance_uid.as_ref()
+    }
+
+    // These methods are noop and should be optimized out
+    fn publish(&self, _event_name: String, _send: Value, _request: Option<&HttpRequest>) {}
+    fn get_search(&self, _aggregate: super::SearchAggregator) {}
+    fn post_search(&self, _aggregate: super::SearchAggregator) {}
+    fn get_similar(&self, _aggregate: super::SimilarAggregator) {}
+    fn post_similar(&self, _aggregate: super::SimilarAggregator) {}
+    fn post_multi_search(&self, _aggregate: super::MultiSearchAggregator) {}
+    fn post_facet_search(&self, _aggregate: super::FacetSearchAggregator) {}
+    fn add_documents(
+        &self,
+        _documents_query: &UpdateDocumentsQuery,
+        _index_creation: bool,
+        _request: &HttpRequest,
+    ) {
+    }
+    fn delete_documents(&self, _kind: DocumentDeletionKind, _request: &HttpRequest) {}
+    fn update_documents(
+        &self,
+        _documents_query: &UpdateDocumentsQuery,
+        _index_creation: bool,
+        _request: &HttpRequest,
+    ) {
+    }
+    fn update_documents_by_function(
+        &self,
+        _documents_query: &DocumentEditionByFunction,
+        _index_creation: bool,
+        _request: &HttpRequest,
+    ) {
+    }
+    fn get_fetch_documents(&self, _documents_query: &DocumentFetchKind, _request: &HttpRequest) {}
+    fn post_fetch_documents(&self, _documents_query: &DocumentFetchKind, _request: &HttpRequest) {}
+}

crates/meilisearch/src/analytics/mod.rs

@@ -0,0 +1,166 @@
+pub mod segment_analytics;
+
+use std::fs;
+use std::path::{Path, PathBuf};
+use std::str::FromStr;
+use std::sync::Arc;
+
+use actix_web::HttpRequest;
+use index_scheduler::IndexScheduler;
+use meilisearch_auth::AuthController;
+use meilisearch_types::InstanceUid;
+use mopa::mopafy;
+use once_cell::sync::Lazy;
+use platform_dirs::AppDirs;
+
+// if the feature analytics is enabled we use the real analytics
+pub type SegmentAnalytics = segment_analytics::SegmentAnalytics;
+
+use crate::Opt;
+
+/// A macro used to quickly define events that don't aggregate or send anything besides an empty event with its name.
+#[macro_export]
+macro_rules! empty_analytics {
+    ($struct_name:ident, $event_name:literal) => {
+        #[derive(Default)]
+        struct $struct_name {}
+
+        impl $crate::analytics::Aggregate for $struct_name {
+            fn event_name(&self) -> &'static str {
+                $event_name
+            }
+
+            fn aggregate(self: Box<Self>, _other: Box<Self>) -> Box<Self> {
+                self
+            }
+
+            fn into_event(self: Box<Self>) -> serde_json::Value {
+                serde_json::json!({})
+            }
+        }
+    };
+}
+
+/// The Meilisearch config dir:
+/// `~/.config/Meilisearch` on *NIX or *BSD.
+/// `~/Library/ApplicationSupport` on macOS.
+/// `%APPDATA` (= `C:\Users%USERNAME%\AppData\Roaming`) on windows.
+static MEILISEARCH_CONFIG_PATH: Lazy<Option<PathBuf>> =
+    Lazy::new(|| AppDirs::new(Some("Meilisearch"), false).map(|appdir| appdir.config_dir));
+
+fn config_user_id_path(db_path: &Path) -> Option<PathBuf> {
+    db_path
+        .canonicalize()
+        .ok()
+        .map(|path| path.join("instance-uid").display().to_string().replace('/', "-"))
+        .zip(MEILISEARCH_CONFIG_PATH.as_ref())
+        .map(|(filename, config_path)| config_path.join(filename.trim_start_matches('-')))
+}
+
+/// Look for the instance-uid in the `data.ms` or in `~/.config/Meilisearch/path-to-db-instance-uid`
+fn find_user_id(db_path: &Path) -> Option<InstanceUid> {
+    fs::read_to_string(db_path.join("instance-uid"))
+        .ok()
+        .or_else(|| fs::read_to_string(config_user_id_path(db_path)?).ok())
+        .and_then(|uid| InstanceUid::from_str(&uid).ok())
+}
+
+#[derive(Copy, Clone, Debug, PartialEq, Eq)]
+pub enum DocumentDeletionKind {
+    PerDocumentId,
+    ClearAll,
+    PerBatch,
+    PerFilter,
+}
+
+#[derive(Copy, Clone, Debug, PartialEq, Eq)]
+pub enum DocumentFetchKind {
+    PerDocumentId { retrieve_vectors: bool },
+    Normal { with_filter: bool, limit: usize, offset: usize, retrieve_vectors: bool },
+}
+
+/// To send an event to segment, your event must be able to aggregate itself with another event of the same type.
+pub trait Aggregate: 'static + mopa::Any + Send {
+    /// The name of the event that will be sent to segment.
+    fn event_name(&self) -> &'static str;
+
+    /// Will be called every time an event has been used twice before segment flushed its buffer.
+    fn aggregate(self: Box<Self>, new: Box<Self>) -> Box<Self>
+    where
+        Self: Sized;
+
+    /// Converts your structure to the final event that'll be sent to segment.
+    fn into_event(self: Box<Self>) -> serde_json::Value;
+}
+
+mopafy!(Aggregate);
+
+/// Helper trait to define multiple aggregates with the same content but a different name.
+/// Commonly used when you must aggregate a search with POST or with GET, for example.
+pub trait AggregateMethod: 'static + Default + Send {
+    fn event_name() -> &'static str;
+}
+
+/// A macro used to quickly define multiple aggregate method with their name
+/// Usage:
+/// ```rust
+/// use meilisearch::aggregate_methods;
+///
+/// aggregate_methods!(
+///     SearchGET => "Documents Searched GET",
+///     SearchPOST => "Documents Searched POST",
+/// );
+/// ```
+#[macro_export]
+macro_rules! aggregate_methods {
+    ($method:ident => $event_name:literal) => {
+        #[derive(Default)]
+        pub struct $method {}
+
+        impl $crate::analytics::AggregateMethod for $method {
+            fn event_name() -> &'static str {
+                $event_name
+            }
+        }
+    };
+    ($($method:ident => $event_name:literal,)+) => {
+        $(
+            aggregate_methods!($method => $event_name);
+        )+
+
+    };
+}
+
+#[derive(Clone)]
+pub struct Analytics {
+    segment: Option<Arc<SegmentAnalytics>>,
+}
+
+impl Analytics {
+    pub async fn new(
+        opt: &Opt,
+        index_scheduler: Arc<IndexScheduler>,
+        auth_controller: Arc<AuthController>,
+    ) -> Self {
+        if opt.no_analytics {
+            Self { segment: None }
+        } else {
+            Self { segment: SegmentAnalytics::new(opt, index_scheduler, auth_controller).await }
+        }
+    }
+
+    pub fn no_analytics() -> Self {
+        Self { segment: None }
+    }
+
+    pub fn instance_uid(&self) -> Option<&InstanceUid> {
+        self.segment.as_ref().map(|segment| segment.instance_uid.as_ref())
+    }
+
+    /// The method used to publish most analytics that do not need to be batched every hours
+    pub fn publish<T: Aggregate>(&self, event: T, request: &HttpRequest) {
+        if let Some(ref segment) = self.segment {
+            let _ = segment.sender.try_send(segment_analytics::Message::new(event, request));
+        }
+    }
+}

crates/meilisearch/src/analytics/segment_analytics.rs

@@ -0,0 +1,484 @@
+use std::any::TypeId;
+use std::collections::{HashMap, HashSet};
+use std::fs;
+use std::path::{Path, PathBuf};
+use std::sync::Arc;
+use std::time::{Duration, Instant};
+
+use actix_web::http::header::USER_AGENT;
+use actix_web::HttpRequest;
+use byte_unit::Byte;
+use index_scheduler::IndexScheduler;
+use meilisearch_auth::{AuthController, AuthFilter};
+use meilisearch_types::features::RuntimeTogglableFeatures;
+use meilisearch_types::InstanceUid;
+use once_cell::sync::Lazy;
+use segment::message::{Identify, Track, User};
+use segment::{AutoBatcher, Batcher, HttpClient};
+use serde::Serialize;
+use serde_json::{json, Value};
+use sysinfo::{Disks, System};
+use time::OffsetDateTime;
+use tokio::select;
+use tokio::sync::mpsc::{self, Receiver, Sender};
+use uuid::Uuid;
+
+use super::{config_user_id_path, Aggregate, MEILISEARCH_CONFIG_PATH};
+use crate::option::{
+    default_http_addr, IndexerOpts, LogMode, MaxMemory, MaxThreads, ScheduleSnapshot,
+};
+use crate::routes::{create_all_stats, Stats};
+use crate::Opt;
+
+const ANALYTICS_HEADER: &str = "X-Meilisearch-Client";
+
+/// Write the instance-uid in the `data.ms` and in `~/.config/MeiliSearch/path-to-db-instance-uid`. Ignore the errors.
+fn write_user_id(db_path: &Path, user_id: &InstanceUid) {
+    let _ = fs::write(db_path.join("instance-uid"), user_id.to_string());
+    if let Some((meilisearch_config_path, user_id_path)) =
+        MEILISEARCH_CONFIG_PATH.as_ref().zip(config_user_id_path(db_path))
+    {
+        let _ = fs::create_dir_all(meilisearch_config_path);
+        let _ = fs::write(user_id_path, user_id.to_string());
+    }
+}
+
+const SEGMENT_API_KEY: &str = "P3FWhhEsJiEDCuEHpmcN9DHcK4hVfBvb";
+
+pub fn extract_user_agents(request: &HttpRequest) -> HashSet<String> {
+    request
+        .headers()
+        .get(ANALYTICS_HEADER)
+        .or_else(|| request.headers().get(USER_AGENT))
+        .and_then(|header| header.to_str().ok())
+        .unwrap_or("unknown")
+        .split(';')
+        .map(str::trim)
+        .map(ToString::to_string)
+        .collect()
+}
+
+pub struct Message {
+    // Since the type_id is solved statically we cannot retrieve it from the Box.
+    // Thus we have to send it in the message directly.
+    type_id: TypeId,
+    // Same for the aggregate function.
+    #[allow(clippy::type_complexity)]
+    aggregator_function: fn(Box<dyn Aggregate>, Box<dyn Aggregate>) -> Option<Box<dyn Aggregate>>,
+    event: Event,
+}
+
+pub struct Event {
+    original: Box<dyn Aggregate>,
+    timestamp: OffsetDateTime,
+    user_agents: HashSet<String>,
+    total: usize,
+}
+
+/// This function should always be called on the same type. If `this` and `other`
+/// aren't the same type the function will do nothing and return `None`.
+fn downcast_aggregate<ConcreteType: Aggregate>(
+    old: Box<dyn Aggregate>,
+    new: Box<dyn Aggregate>,
+) -> Option<Box<dyn Aggregate>> {
+    if old.is::<ConcreteType>() && new.is::<ConcreteType>() {
+        // Both the two following lines cannot fail, but just to be sure we don't crash, we're still avoiding unwrapping
+        let this = old.downcast::<ConcreteType>().ok()?;
+        let other = new.downcast::<ConcreteType>().ok()?;
+        Some(ConcreteType::aggregate(this, other))
+    } else {
+        None
+    }
+}
+
+impl Message {
+    pub fn new<T: Aggregate>(event: T, request: &HttpRequest) -> Self {
+        Self {
+            type_id: TypeId::of::<T>(),
+            event: Event {
+                original: Box::new(event),
+                timestamp: OffsetDateTime::now_utc(),
+                user_agents: extract_user_agents(request),
+                total: 1,
+            },
+            aggregator_function: downcast_aggregate::<T>,
+        }
+    }
+}
+
+pub struct SegmentAnalytics {
+    pub instance_uid: InstanceUid,
+    pub user: User,
+    pub sender: Sender<Message>,
+}
+
+impl SegmentAnalytics {
+    #[allow(clippy::new_ret_no_self)]
+    pub async fn new(
+        opt: &Opt,
+        index_scheduler: Arc<IndexScheduler>,
+        auth_controller: Arc<AuthController>,
+    ) -> Option<Arc<Self>> {
+        let instance_uid = super::find_user_id(&opt.db_path);
+        let first_time_run = instance_uid.is_none();
+        let instance_uid = instance_uid.unwrap_or_else(Uuid::new_v4);
+        write_user_id(&opt.db_path, &instance_uid);
+
+        let client = reqwest::Client::builder().connect_timeout(Duration::from_secs(10)).build();
+
+        // if reqwest throws an error we won't be able to send analytics
+        if client.is_err() {
+            return None;
+        }
+
+        let client =
+            HttpClient::new(client.unwrap(), "https://telemetry.meilisearch.com".to_string());
+        let user = User::UserId { user_id: instance_uid.to_string() };
+        let mut batcher = AutoBatcher::new(client, Batcher::new(None), SEGMENT_API_KEY.to_string());
+
+        // If Meilisearch is Launched for the first time:
+        // 1. Send an event Launched associated to the user `total_launch`.
+        // 2. Batch an event Launched with the real instance-id and send it in one hour.
+        if first_time_run {
+            let _ = batcher
+                .push(Track {
+                    user: User::UserId { user_id: "total_launch".to_string() },
+                    event: "Launched".to_string(),
+                    ..Default::default()
+                })
+                .await;
+            let _ = batcher.flush().await;
+            let _ = batcher
+                .push(Track {
+                    user: user.clone(),
+                    event: "Launched".to_string(),
+                    ..Default::default()
+                })
+                .await;
+        }
+
+        let (sender, inbox) = mpsc::channel(100); // How many analytics can we bufferize
+
+        let segment = Box::new(Segment {
+            inbox,
+            user: user.clone(),
+            opt: opt.clone(),
+            batcher,
+            events: HashMap::new(),
+        });
+        tokio::spawn(segment.run(index_scheduler.clone(), auth_controller.clone()));
+
+        let this = Self { instance_uid, sender, user: user.clone() };
+
+        Some(Arc::new(this))
+    }
+}
+
+/// This structure represent the `infos` field we send in the analytics.
+/// It's quite close to the `Opt` structure except all sensitive informations
+/// have been simplified to a boolean.
+/// It's send as-is in amplitude thus you should never update a name of the
+/// struct without the approval of the PM.
+#[derive(Debug, Clone, Serialize)]
+struct Infos {
+    env: String,
+    experimental_contains_filter: bool,
+    experimental_vector_store: bool,
+    experimental_enable_metrics: bool,
+    experimental_edit_documents_by_function: bool,
+    experimental_search_queue_size: usize,
+    experimental_drop_search_after: usize,
+    experimental_nb_searches_per_core: usize,
+    experimental_logs_mode: LogMode,
+    experimental_replication_parameters: bool,
+    experimental_enable_logs_route: bool,
+    experimental_reduce_indexing_memory_usage: bool,
+    experimental_max_number_of_batched_tasks: usize,
+    gpu_enabled: bool,
+    db_path: bool,
+    import_dump: bool,
+    dump_dir: bool,
+    ignore_missing_dump: bool,
+    ignore_dump_if_db_exists: bool,
+    import_snapshot: bool,
+    schedule_snapshot: Option<u64>,
+    snapshot_dir: bool,
+    ignore_missing_snapshot: bool,
+    ignore_snapshot_if_db_exists: bool,
+    http_addr: bool,
+    http_payload_size_limit: Byte,
+    task_queue_webhook: bool,
+    task_webhook_authorization_header: bool,
+    log_level: String,
+    max_indexing_memory: MaxMemory,
+    max_indexing_threads: MaxThreads,
+    with_configuration_file: bool,
+    ssl_auth_path: bool,
+    ssl_cert_path: bool,
+    ssl_key_path: bool,
+    ssl_ocsp_path: bool,
+    ssl_require_auth: bool,
+    ssl_resumption: bool,
+    ssl_tickets: bool,
+}
+
+impl Infos {
+    pub fn new(options: Opt, features: RuntimeTogglableFeatures) -> Self {
+        // We wants to decompose this whole struct by hand to be sure we don't forget
+        // to add analytics when we add a field in the Opt.
+        // Thus we must not insert `..` at the end.
+        let Opt {
+            db_path,
+            experimental_contains_filter,
+            experimental_enable_metrics,
+            experimental_search_queue_size,
+            experimental_drop_search_after,
+            experimental_nb_searches_per_core,
+            experimental_logs_mode,
+            experimental_replication_parameters,
+            experimental_enable_logs_route,
+            experimental_reduce_indexing_memory_usage,
+            experimental_max_number_of_batched_tasks,
+            http_addr,
+            master_key: _,
+            env,
+            task_webhook_url,
+            task_webhook_authorization_header,
+            max_index_size: _,
+            max_task_db_size: _,
+            http_payload_size_limit,
+            ssl_cert_path,
+            ssl_key_path,
+            ssl_auth_path,
+            ssl_ocsp_path,
+            ssl_require_auth,
+            ssl_resumption,
+            ssl_tickets,
+            import_snapshot,
+            ignore_missing_snapshot,
+            ignore_snapshot_if_db_exists,
+            snapshot_dir,
+            schedule_snapshot,
+            import_dump,
+            ignore_missing_dump,
+            ignore_dump_if_db_exists,
+            dump_dir,
+            log_level,
+            indexer_options,
+            config_file_path,
+            no_analytics: _,
+        } = options;
+
+        let schedule_snapshot = match schedule_snapshot {
+            ScheduleSnapshot::Disabled => None,
+            ScheduleSnapshot::Enabled(interval) => Some(interval),
+        };
+
+        let IndexerOpts { max_indexing_memory, max_indexing_threads, skip_index_budget: _ } =
+            indexer_options;
+
+        let RuntimeTogglableFeatures {
+            vector_store,
+            metrics,
+            logs_route,
+            edit_documents_by_function,
+            contains_filter,
+        } = features;
+
+        // We're going to override every sensible information.
+        // We consider information sensible if it contains a path, an address, or a key.
+        Self {
+            env,
+            experimental_contains_filter: experimental_contains_filter | contains_filter,
+            experimental_vector_store: vector_store,
+            experimental_edit_documents_by_function: edit_documents_by_function,
+            experimental_enable_metrics: experimental_enable_metrics | metrics,
+            experimental_search_queue_size,
+            experimental_drop_search_after: experimental_drop_search_after.into(),
+            experimental_nb_searches_per_core: experimental_nb_searches_per_core.into(),
+            experimental_logs_mode,
+            experimental_replication_parameters,
+            experimental_enable_logs_route: experimental_enable_logs_route | logs_route,
+            experimental_reduce_indexing_memory_usage,
+            gpu_enabled: meilisearch_types::milli::vector::is_cuda_enabled(),
+            db_path: db_path != PathBuf::from("./data.ms"),
+            import_dump: import_dump.is_some(),
+            dump_dir: dump_dir != PathBuf::from("dumps/"),
+            ignore_missing_dump,
+            ignore_dump_if_db_exists,
+            import_snapshot: import_snapshot.is_some(),
+            schedule_snapshot,
+            snapshot_dir: snapshot_dir != PathBuf::from("snapshots/"),
+            ignore_missing_snapshot,
+            ignore_snapshot_if_db_exists,
+            http_addr: http_addr != default_http_addr(),
+            http_payload_size_limit,
+            experimental_max_number_of_batched_tasks,
+            task_queue_webhook: task_webhook_url.is_some(),
+            task_webhook_authorization_header: task_webhook_authorization_header.is_some(),
+            log_level: log_level.to_string(),
+            max_indexing_memory,
+            max_indexing_threads,
+            with_configuration_file: config_file_path.is_some(),
+            ssl_auth_path: ssl_auth_path.is_some(),
+            ssl_cert_path: ssl_cert_path.is_some(),
+            ssl_key_path: ssl_key_path.is_some(),
+            ssl_ocsp_path: ssl_ocsp_path.is_some(),
+            ssl_require_auth,
+            ssl_resumption,
+            ssl_tickets,
+        }
+    }
+}
+
+pub struct Segment {
+    inbox: Receiver<Message>,
+    user: User,
+    opt: Opt,
+    batcher: AutoBatcher,
+    events: HashMap<TypeId, Event>,
+}
+
+impl Segment {
+    fn compute_traits(opt: &Opt, stats: Stats, features: RuntimeTogglableFeatures) -> Value {
+        static FIRST_START_TIMESTAMP: Lazy<Instant> = Lazy::new(Instant::now);
+        static SYSTEM: Lazy<Value> = Lazy::new(|| {
+            let disks = Disks::new_with_refreshed_list();
+            let mut sys = System::new_all();
+            sys.refresh_all();
+            let kernel_version = System::kernel_version()
+                .and_then(|k| k.split_once('-').map(|(k, _)| k.to_string()));
+            json!({
+                    "distribution": System::name(),
+                    "kernel_version": kernel_version,
+                    "cores": sys.cpus().len(),
+                    "ram_size": sys.total_memory(),
+                    "disk_size": disks.iter().map(|disk| disk.total_space()).max(),
+                    "server_provider": std::env::var("MEILI_SERVER_PROVIDER").ok(),
+            })
+        });
+        let number_of_documents =
+            stats.indexes.values().map(|index| index.number_of_documents).collect::<Vec<u64>>();
+
+        json!({
+            "start_since_days": FIRST_START_TIMESTAMP.elapsed().as_secs() / (60 * 60 * 24), // one day
+            "system": *SYSTEM,
+            "stats": {
+                "database_size": stats.database_size,
+                "indexes_number": stats.indexes.len(),
+                "documents_number": number_of_documents,
+            },
+            "infos": Infos::new(opt.clone(), features),
+        })
+    }
+
+    async fn run(
+        mut self,
+        index_scheduler: Arc<IndexScheduler>,
+        auth_controller: Arc<AuthController>,
+    ) {
+        const INTERVAL: Duration = Duration::from_secs(60 * 60); // one hour
+                                                                 // The first batch must be sent after one hour.
+        let mut interval =
+            tokio::time::interval_at(tokio::time::Instant::now() + INTERVAL, INTERVAL);
+
+        loop {
+            select! {
+                _ = interval.tick() => {
+                    self.tick(index_scheduler.clone(), auth_controller.clone()).await;
+                },
+                Some(msg) = self.inbox.recv() => {
+                    self.handle_msg(msg);
+               }
+            }
+        }
+    }
+
+    fn handle_msg(&mut self, Message { type_id, aggregator_function, event }: Message) {
+        let new_event = match self.events.remove(&type_id) {
+            Some(old) => {
+                // The function should never fail since we retrieved the corresponding TypeId in the map. But in the unfortunate
+                // case it could happens we're going to silently ignore the error
+                let Some(original) = (aggregator_function)(old.original, event.original) else {
+                    return;
+                };
+                Event {
+                    original,
+                    // We always want to return the FIRST timestamp ever encountered
+                    timestamp: old.timestamp,
+                    user_agents: old.user_agents.union(&event.user_agents).cloned().collect(),
+                    total: old.total.saturating_add(event.total),
+                }
+            }
+            None => event,
+        };
+        self.events.insert(type_id, new_event);
+    }
+
+    async fn tick(
+        &mut self,
+        index_scheduler: Arc<IndexScheduler>,
+        auth_controller: Arc<AuthController>,
+    ) {
+        if let Ok(stats) = create_all_stats(
+            index_scheduler.clone().into(),
+            auth_controller.into(),
+            &AuthFilter::default(),
+        ) {
+            // Replace the version number with the prototype name if any.
+            let version = if let Some(prototype) = build_info::DescribeResult::from_build()
+                .and_then(|describe| describe.as_prototype())
+            {
+                prototype
+            } else {
+                env!("CARGO_PKG_VERSION")
+            };
+
+            let _ = self
+                .batcher
+                .push(Identify {
+                    context: Some(json!({
+                        "app": {
+                            "version": version.to_string(),
+                        },
+                    })),
+                    user: self.user.clone(),
+                    traits: Self::compute_traits(
+                        &self.opt,
+                        stats,
+                        index_scheduler.features().runtime_features(),
+                    ),
+                    ..Default::default()
+                })
+                .await;
+        }
+
+        // We empty the list of events
+        let events = std::mem::take(&mut self.events);
+
+        for (_, event) in events {
+            let Event { original, timestamp, user_agents, total } = event;
+            let name = original.event_name();
+            let mut properties = original.into_event();
+            if properties["user-agent"].is_null() {
+                properties["user-agent"] = json!(user_agents);
+            };
+            if properties["requests"]["total_received"].is_null() {
+                properties["requests"]["total_received"] = total.into();
+            };
+
+            let _ = self
+                .batcher
+                .push(Track {
+                    user: self.user.clone(),
+                    event: name.to_string(),
+                    properties,
+                    timestamp: Some(timestamp),
+                    ..Default::default()
+                })
+                .await;
+        }
+
+        let _ = self.batcher.flush().await;
+    }
+}

crates/meilisearch/src/error.rs

@@ -0,0 +1,227 @@
+use actix_web as aweb;
+use aweb::error::{JsonPayloadError, QueryPayloadError};
+use byte_unit::{Byte, UnitType};
+use meilisearch_types::document_formats::{DocumentFormatError, PayloadType};
+use meilisearch_types::error::{Code, ErrorCode, ResponseError};
+use meilisearch_types::index_uid::{IndexUid, IndexUidFormatError};
+use meilisearch_types::milli::OrderBy;
+use serde_json::Value;
+use tokio::task::JoinError;
+
+#[derive(Debug, thiserror::Error)]
+pub enum MeilisearchHttpError {
+    #[error("A Content-Type header is missing. Accepted values for the Content-Type header are: {}",
+            .0.iter().map(|s| format!("`{}`", s)).collect::<Vec<_>>().join(", "))]
+    MissingContentType(Vec<String>),
+    #[error("The `/logs/stream` route is currently in use by someone else.")]
+    AlreadyUsedLogRoute,
+    #[error("The Content-Type `{0}` does not support the use of a csv delimiter. The csv delimiter can only be used with the Content-Type `text/csv`.")]
+    CsvDelimiterWithWrongContentType(String),
+    #[error(
+        "The Content-Type `{0}` is invalid. Accepted values for the Content-Type header are: {}",
+        .1.iter().map(|s| format!("`{}`", s)).collect::<Vec<_>>().join(", ")
+    )]
+    InvalidContentType(String, Vec<String>),
+    #[error("Document `{0}` not found.")]
+    DocumentNotFound(String),
+    #[error("Sending an empty filter is forbidden.")]
+    EmptyFilter,
+    #[error("Invalid syntax for the filter parameter: `expected {}, found: {1}`.", .0.join(", "))]
+    InvalidExpression(&'static [&'static str], Value),
+    #[error("Using `federationOptions` is not allowed in a non-federated search.\n - Hint: remove `federationOptions` from query #{0} or add `federation` to the request.")]
+    FederationOptionsInNonFederatedRequest(usize),
+    #[error("Inside `.queries[{0}]`: Using pagination options is not allowed in federated queries.\n - Hint: remove `{1}` from query #{0} or remove `federation` from the request\n - Hint: pass `federation.limit` and `federation.offset` for pagination in federated search")]
+    PaginationInFederatedQuery(usize, &'static str),
+    #[error("Inside `.queries[{0}]`: Using facet options is not allowed in federated queries.\n - Hint: remove `facets` from query #{0} or remove `federation` from the request\n - Hint: pass `federation.facetsByIndex.{1}: {2:?}` for facets in federated search")]
+    FacetsInFederatedQuery(usize, String, Vec<String>),
+    #[error("Inconsistent order for values in facet `{facet}`: index `{previous_uid}` orders {previous_facet_order}, but index `{current_uid}` orders {index_facet_order}.\n - Hint: Remove `federation.mergeFacets` or change `faceting.sortFacetValuesBy` to be consistent in settings.")]
+    InconsistentFacetOrder {
+        facet: String,
+        previous_facet_order: OrderBy,
+        previous_uid: String,
+        index_facet_order: OrderBy,
+        current_uid: String,
+    },
+    #[error("A {0} payload is missing.")]
+    MissingPayload(PayloadType),
+    #[error("Too many search requests running at the same time: {0}. Retry after 10s.")]
+    TooManySearchRequests(usize),
+    #[error("Internal error: Search limiter is down.")]
+    SearchLimiterIsDown,
+    #[error("The provided payload reached the size limit. The maximum accepted payload size is {}.",  Byte::from_u64(*.0 as u64).get_appropriate_unit(UnitType::Binary))]
+    PayloadTooLarge(usize),
+    #[error("Two indexes must be given for each swap. The list `[{}]` contains {} indexes.",
+        .0.iter().map(|uid| format!("\"{uid}\"")).collect::<Vec<_>>().join(", "), .0.len()
+    )]
+    SwapIndexPayloadWrongLength(Vec<IndexUid>),
+    #[error(transparent)]
+    IndexUid(#[from] IndexUidFormatError),
+    #[error(transparent)]
+    SerdeJson(#[from] serde_json::Error),
+    #[error(transparent)]
+    HeedError(#[from] meilisearch_types::heed::Error),
+    #[error(transparent)]
+    IndexScheduler(#[from] index_scheduler::Error),
+    #[error(transparent)]
+    Milli(#[from] meilisearch_types::milli::Error),
+    #[error(transparent)]
+    Payload(#[from] PayloadError),
+    #[error(transparent)]
+    FileStore(#[from] file_store::Error),
+    #[error(transparent)]
+    DocumentFormat(#[from] DocumentFormatError),
+    #[error(transparent)]
+    Join(#[from] JoinError),
+    #[error("Invalid request: missing `hybrid` parameter when `vector` is present.")]
+    MissingSearchHybrid,
+}
+
+impl ErrorCode for MeilisearchHttpError {
+    fn error_code(&self) -> Code {
+        match self {
+            MeilisearchHttpError::MissingContentType(_) => Code::MissingContentType,
+            MeilisearchHttpError::AlreadyUsedLogRoute => Code::BadRequest,
+            MeilisearchHttpError::CsvDelimiterWithWrongContentType(_) => Code::InvalidContentType,
+            MeilisearchHttpError::MissingPayload(_) => Code::MissingPayload,
+            MeilisearchHttpError::InvalidContentType(_, _) => Code::InvalidContentType,
+            MeilisearchHttpError::DocumentNotFound(_) => Code::DocumentNotFound,
+            MeilisearchHttpError::EmptyFilter => Code::InvalidDocumentFilter,
+            MeilisearchHttpError::InvalidExpression(_, _) => Code::InvalidSearchFilter,
+            MeilisearchHttpError::PayloadTooLarge(_) => Code::PayloadTooLarge,
+            MeilisearchHttpError::TooManySearchRequests(_) => Code::TooManySearchRequests,
+            MeilisearchHttpError::SearchLimiterIsDown => Code::Internal,
+            MeilisearchHttpError::SwapIndexPayloadWrongLength(_) => Code::InvalidSwapIndexes,
+            MeilisearchHttpError::IndexUid(e) => e.error_code(),
+            MeilisearchHttpError::SerdeJson(_) => Code::Internal,
+            MeilisearchHttpError::HeedError(_) => Code::Internal,
+            MeilisearchHttpError::IndexScheduler(e) => e.error_code(),
+            MeilisearchHttpError::Milli(e) => e.error_code(),
+            MeilisearchHttpError::Payload(e) => e.error_code(),
+            MeilisearchHttpError::FileStore(_) => Code::Internal,
+            MeilisearchHttpError::DocumentFormat(e) => e.error_code(),
+            MeilisearchHttpError::Join(_) => Code::Internal,
+            MeilisearchHttpError::MissingSearchHybrid => Code::MissingSearchHybrid,
+            MeilisearchHttpError::FederationOptionsInNonFederatedRequest(_) => {
+                Code::InvalidMultiSearchFederationOptions
+            }
+            MeilisearchHttpError::PaginationInFederatedQuery(_, _) => {
+                Code::InvalidMultiSearchQueryPagination
+            }
+            MeilisearchHttpError::FacetsInFederatedQuery(..) => Code::InvalidMultiSearchQueryFacets,
+            MeilisearchHttpError::InconsistentFacetOrder { .. } => {
+                Code::InvalidMultiSearchFacetOrder
+            }
+        }
+    }
+}
+
+impl From<MeilisearchHttpError> for aweb::Error {
+    fn from(other: MeilisearchHttpError) -> Self {
+        aweb::Error::from(ResponseError::from(other))
+    }
+}
+
+impl From<aweb::error::PayloadError> for MeilisearchHttpError {
+    fn from(error: aweb::error::PayloadError) -> Self {
+        match error {
+            aweb::error::PayloadError::Incomplete(_) => MeilisearchHttpError::Payload(
+                PayloadError::Payload(ActixPayloadError::IncompleteError),
+            ),
+            _ => MeilisearchHttpError::Payload(PayloadError::Payload(
+                ActixPayloadError::OtherError(error),
+            )),
+        }
+    }
+}
+
+#[derive(Debug, thiserror::Error)]
+pub enum ActixPayloadError {
+    #[error("The provided payload is incomplete and cannot be parsed")]
+    IncompleteError,
+    #[error(transparent)]
+    OtherError(aweb::error::PayloadError),
+}
+
+#[derive(Debug, thiserror::Error)]
+pub enum PayloadError {
+    #[error(transparent)]
+    Payload(ActixPayloadError),
+    #[error(transparent)]
+    Json(JsonPayloadError),
+    #[error(transparent)]
+    Query(QueryPayloadError),
+    #[error("The json payload provided is malformed. `{0}`.")]
+    MalformedPayload(serde_json::error::Error),
+    #[error("A json payload is missing.")]
+    MissingPayload,
+    #[error("Error while receiving the playload. `{0}`.")]
+    ReceivePayload(Box<dyn std::error::Error + Send + Sync + 'static>),
+}
+
+impl ErrorCode for PayloadError {
+    fn error_code(&self) -> Code {
+        match self {
+            PayloadError::Payload(e) => match e {
+                ActixPayloadError::IncompleteError => Code::BadRequest,
+                ActixPayloadError::OtherError(error) => match error {
+                    aweb::error::PayloadError::EncodingCorrupted => Code::Internal,
+                    aweb::error::PayloadError::Overflow => Code::PayloadTooLarge,
+                    aweb::error::PayloadError::UnknownLength => Code::Internal,
+                    aweb::error::PayloadError::Http2Payload(_) => Code::Internal,
+                    aweb::error::PayloadError::Io(_) => Code::Internal,
+                    _ => todo!(),
+                },
+            },
+            PayloadError::Json(err) => match err {
+                JsonPayloadError::Overflow { .. } => Code::PayloadTooLarge,
+                JsonPayloadError::ContentType => Code::UnsupportedMediaType,
+                JsonPayloadError::Payload(aweb::error::PayloadError::Overflow) => {
+                    Code::PayloadTooLarge
+                }
+                JsonPayloadError::Payload(_) => Code::BadRequest,
+                JsonPayloadError::Deserialize(_) => Code::BadRequest,
+                JsonPayloadError::Serialize(_) => Code::Internal,
+                _ => Code::Internal,
+            },
+            PayloadError::Query(err) => match err {
+                QueryPayloadError::Deserialize(_) => Code::BadRequest,
+                _ => Code::Internal,
+            },
+            PayloadError::MissingPayload => Code::MissingPayload,
+            PayloadError::MalformedPayload(_) => Code::MalformedPayload,
+            PayloadError::ReceivePayload(_) => Code::Internal,
+        }
+    }
+}
+
+impl From<JsonPayloadError> for PayloadError {
+    fn from(other: JsonPayloadError) -> Self {
+        match other {
+            JsonPayloadError::Deserialize(e)
+                if e.classify() == serde_json::error::Category::Eof
+                    && e.line() == 1
+                    && e.column() == 0 =>
+            {
+                Self::MissingPayload
+            }
+            JsonPayloadError::Deserialize(e)
+                if e.classify() != serde_json::error::Category::Data =>
+            {
+                Self::MalformedPayload(e)
+            }
+            _ => Self::Json(other),
+        }
+    }
+}
+
+impl From<QueryPayloadError> for PayloadError {
+    fn from(other: QueryPayloadError) -> Self {
+        Self::Query(other)
+    }
+}
+
+impl From<PayloadError> for aweb::Error {
+    fn from(other: PayloadError) -> Self {
+        aweb::Error::from(ResponseError::from(other))
+    }
+}

crates/meilisearch/src/extractors/authentication/error.rs

@@ -0,0 +1,25 @@
+use meilisearch_types::error::{Code, ErrorCode};
+
+#[derive(Debug, thiserror::Error)]
+pub enum AuthenticationError {
+    #[error("The Authorization header is missing. It must use the bearer authorization method.")]
+    MissingAuthorizationHeader,
+    #[error("The provided API key is invalid.")]
+    InvalidToken,
+    // Triggered on configuration error.
+    #[error("An internal error has occurred. `Irretrievable state`.")]
+    IrretrievableState,
+    #[error("Meilisearch is running without a master key. To access this API endpoint, you must have set a master key at launch.")]
+    MissingMasterKey,
+}
+
+impl ErrorCode for AuthenticationError {
+    fn error_code(&self) -> Code {
+        match self {
+            AuthenticationError::MissingAuthorizationHeader => Code::MissingAuthorizationHeader,
+            AuthenticationError::InvalidToken => Code::InvalidApiKey,
+            AuthenticationError::IrretrievableState => Code::Internal,
+            AuthenticationError::MissingMasterKey => Code::MissingMasterKey,
+        }
+    }
+}

crates/meilisearch/src/extractors/authentication/mod.rs

@@ -0,0 +1,341 @@
+mod error;
+
+use std::marker::PhantomData;
+use std::ops::Deref;
+use std::pin::Pin;
+
+use actix_web::web::Data;
+use actix_web::FromRequest;
+pub use error::AuthenticationError;
+use futures::future::err;
+use futures::Future;
+use meilisearch_auth::{AuthController, AuthFilter};
+use meilisearch_types::error::{Code, ResponseError};
+
+use self::policies::AuthError;
+
+pub struct GuardedData<P, D> {
+    data: D,
+    filters: AuthFilter,
+    _marker: PhantomData<P>,
+}
+
+impl<P, D> GuardedData<P, D> {
+    pub fn filters(&self) -> &AuthFilter {
+        &self.filters
+    }
+
+    async fn auth_bearer(
+        auth: Data<AuthController>,
+        token: String,
+        index: Option<String>,
+        data: Option<D>,
+    ) -> Result<Self, ResponseError>
+    where
+        P: Policy + 'static,
+    {
+        let missing_master_key = auth.get_master_key().is_none();
+
+        match Self::authenticate(auth, token, index).await? {
+            Ok(filters) => match data {
+                Some(data) => Ok(Self { data, filters, _marker: PhantomData }),
+                None => Err(AuthenticationError::IrretrievableState.into()),
+            },
+            Err(_) if missing_master_key => Err(AuthenticationError::MissingMasterKey.into()),
+            Err(e) => Err(ResponseError::from_msg(e.to_string(), Code::InvalidApiKey)),
+        }
+    }
+
+    async fn auth_token(auth: Data<AuthController>, data: Option<D>) -> Result<Self, ResponseError>
+    where
+        P: Policy + 'static,
+    {
+        let missing_master_key = auth.get_master_key().is_none();
+
+        match Self::authenticate(auth, String::new(), None).await? {
+            Ok(filters) => match data {
+                Some(data) => Ok(Self { data, filters, _marker: PhantomData }),
+                None => Err(AuthenticationError::IrretrievableState.into()),
+            },
+            Err(_) if missing_master_key => Err(AuthenticationError::MissingMasterKey.into()),
+            Err(_) => Err(AuthenticationError::MissingAuthorizationHeader.into()),
+        }
+    }
+
+    async fn authenticate(
+        auth: Data<AuthController>,
+        token: String,
+        index: Option<String>,
+    ) -> Result<Result<AuthFilter, AuthError>, ResponseError>
+    where
+        P: Policy + 'static,
+    {
+        tokio::task::spawn_blocking(move || P::authenticate(auth, token.as_ref(), index.as_deref()))
+            .await
+            .map_err(|e| ResponseError::from_msg(e.to_string(), Code::Internal))
+    }
+}
+
+impl<P, D> Deref for GuardedData<P, D> {
+    type Target = D;
+
+    fn deref(&self) -> &Self::Target {
+        &self.data
+    }
+}
+
+impl<P: Policy + 'static, D: 'static + Clone> FromRequest for GuardedData<P, D> {
+    type Error = ResponseError;
+
+    type Future = Pin<Box<dyn Future<Output = Result<Self, Self::Error>>>>;
+
+    fn from_request(
+        req: &actix_web::HttpRequest,
+        _payload: &mut actix_web::dev::Payload,
+    ) -> Self::Future {
+        match req.app_data::<Data<AuthController>>().cloned() {
+            Some(auth) => match req
+                .headers()
+                .get("Authorization")
+                .map(|type_token| type_token.to_str().unwrap_or_default().splitn(2, ' '))
+            {
+                Some(mut type_token) => match type_token.next() {
+                    Some("Bearer") => {
+                        // TODO: find a less hardcoded way?
+                        let index = req.match_info().get("index_uid");
+                        match type_token.next() {
+                            Some(token) => Box::pin(Self::auth_bearer(
+                                auth,
+                                token.to_string(),
+                                index.map(String::from),
+                                req.app_data::<D>().cloned(),
+                            )),
+                            None => Box::pin(err(AuthenticationError::InvalidToken.into())),
+                        }
+                    }
+                    _otherwise => {
+                        Box::pin(err(AuthenticationError::MissingAuthorizationHeader.into()))
+                    }
+                },
+                None => Box::pin(Self::auth_token(auth, req.app_data::<D>().cloned())),
+            },
+            None => Box::pin(err(AuthenticationError::IrretrievableState.into())),
+        }
+    }
+}
+
+pub trait Policy {
+    fn authenticate(
+        auth: Data<AuthController>,
+        token: &str,
+        index: Option<&str>,
+    ) -> Result<AuthFilter, policies::AuthError>;
+}
+
+pub mod policies {
+    use actix_web::web::Data;
+    use jsonwebtoken::{decode, Algorithm, DecodingKey, Validation};
+    use meilisearch_auth::{AuthController, AuthFilter, SearchRules};
+    use meilisearch_types::error::{Code, ErrorCode};
+    // reexport actions in policies in order to be used in routes configuration.
+    pub use meilisearch_types::keys::{actions, Action};
+    use serde::{Deserialize, Serialize};
+    use time::OffsetDateTime;
+    use uuid::Uuid;
+
+    use crate::extractors::authentication::Policy;
+
+    enum TenantTokenOutcome {
+        NotATenantToken,
+        Valid(Uuid, SearchRules),
+    }
+
+    #[derive(thiserror::Error, Debug)]
+    pub enum AuthError {
+        #[error("Tenant token expired. Was valid up to `{exp}` and we're now `{now}`.")]
+        ExpiredTenantToken { exp: i64, now: i64 },
+        #[error("The provided API key is invalid.")]
+        InvalidApiKey,
+        #[error("The provided tenant token cannot acces the index `{index}`, allowed indexes are {allowed:?}.")]
+        TenantTokenAccessingnUnauthorizedIndex { index: String, allowed: Vec<String> },
+        #[error(
+            "The API key used to generate this tenant token cannot acces the index `{index}`."
+        )]
+        TenantTokenApiKeyAccessingnUnauthorizedIndex { index: String },
+        #[error(
+            "The API key cannot acces the index `{index}`, authorized indexes are {allowed:?}."
+        )]
+        ApiKeyAccessingnUnauthorizedIndex { index: String, allowed: Vec<String> },
+        #[error("The provided tenant token is invalid.")]
+        InvalidTenantToken,
+        #[error("Could not decode tenant token, {0}.")]
+        CouldNotDecodeTenantToken(jsonwebtoken::errors::Error),
+        #[error("Invalid action `{0}`.")]
+        InternalInvalidAction(u8),
+    }
+
+    impl From<jsonwebtoken::errors::Error> for AuthError {
+        fn from(error: jsonwebtoken::errors::Error) -> Self {
+            use jsonwebtoken::errors::ErrorKind;
+
+            match error.kind() {
+                ErrorKind::InvalidToken => AuthError::InvalidTenantToken,
+                _ => AuthError::CouldNotDecodeTenantToken(error),
+            }
+        }
+    }
+
+    impl ErrorCode for AuthError {
+        fn error_code(&self) -> Code {
+            match self {
+                AuthError::InternalInvalidAction(_) => Code::Internal,
+                _ => Code::InvalidApiKey,
+            }
+        }
+    }
+
+    fn tenant_token_validation() -> Validation {
+        let mut validation = Validation::default();
+        validation.validate_exp = false;
+        validation.required_spec_claims.remove("exp");
+        validation.algorithms = vec![Algorithm::HS256, Algorithm::HS384, Algorithm::HS512];
+        validation
+    }
+
+    /// Extracts the key id used to sign the payload, without performing any validation.
+    fn extract_key_id(token: &str) -> Result<Uuid, AuthError> {
+        let mut validation = tenant_token_validation();
+        validation.insecure_disable_signature_validation();
+        let dummy_key = DecodingKey::from_secret(b"secret");
+        let token_data = decode::<Claims>(token, &dummy_key, &validation)?;
+
+        // get token fields without validating it.
+        let Claims { api_key_uid, .. } = token_data.claims;
+        Ok(api_key_uid)
+    }
+
+    fn is_keys_action(action: u8) -> bool {
+        use actions::*;
+        matches!(action, KEYS_GET | KEYS_CREATE | KEYS_UPDATE | KEYS_DELETE)
+    }
+
+    pub struct ActionPolicy<const A: u8>;
+
+    impl<const A: u8> Policy for ActionPolicy<A> {
+        /// Attempts to grant authentication from a bearer token (that can be a tenant token or an API key), the requested Action,
+        /// and a list of requested indexes.
+        ///
+        /// If the bearer token is not allowed for the specified indexes and action, returns `None`.
+        /// Otherwise, returns an object containing the generated permissions: the search filters to add to a search, and the list of allowed indexes
+        /// (that may contain more indexes than requested).
+        fn authenticate(
+            auth: Data<AuthController>,
+            token: &str,
+            index: Option<&str>,
+        ) -> Result<AuthFilter, AuthError> {
+            // authenticate if token is the master key.
+            // Without a master key, all routes are accessible except the key-related routes.
+            if auth.get_master_key().map_or_else(|| !is_keys_action(A), |mk| mk == token) {
+                return Ok(AuthFilter::default());
+            }
+
+            let (key_uuid, search_rules) =
+                match ActionPolicy::<A>::authenticate_tenant_token(&auth, token) {
+                    Ok(TenantTokenOutcome::Valid(key_uuid, search_rules)) => {
+                        (key_uuid, Some(search_rules))
+                    }
+                    Ok(TenantTokenOutcome::NotATenantToken)
+                    | Err(AuthError::InvalidTenantToken) => (
+                        auth.get_optional_uid_from_encoded_key(token.as_bytes())
+                            .map_err(|_e| AuthError::InvalidApiKey)?
+                            .ok_or(AuthError::InvalidApiKey)?,
+                        None,
+                    ),
+                    Err(e) => return Err(e),
+                };
+
+            // check that the indexes are allowed
+            let action = Action::from_repr(A).ok_or(AuthError::InternalInvalidAction(A))?;
+            let auth_filter = auth
+                .get_key_filters(key_uuid, search_rules)
+                .map_err(|_e| AuthError::InvalidApiKey)?;
+
+            // First check if the index is authorized in the tenant token, this is a public
+            // information, we can return a nice error message.
+            if let Some(index) = index {
+                if !auth_filter.tenant_token_is_index_authorized(index) {
+                    return Err(AuthError::TenantTokenAccessingnUnauthorizedIndex {
+                        index: index.to_string(),
+                        allowed: auth_filter.tenant_token_list_index_authorized(),
+                    });
+                }
+                if !auth_filter.api_key_is_index_authorized(index) {
+                    if auth_filter.is_tenant_token() {
+                        // If the error comes from a tenant token we cannot share the list
+                        // of authorized indexes in the API key. This is not public information.
+                        return Err(AuthError::TenantTokenApiKeyAccessingnUnauthorizedIndex {
+                            index: index.to_string(),
+                        });
+                    } else {
+                        // Otherwise we can share the list
+                        // of authorized indexes in the API key.
+                        return Err(AuthError::ApiKeyAccessingnUnauthorizedIndex {
+                            index: index.to_string(),
+                            allowed: auth_filter.api_key_list_index_authorized(),
+                        });
+                    }
+                }
+            }
+            if auth.is_key_authorized(key_uuid, action, index).unwrap_or(false) {
+                return Ok(auth_filter);
+            }
+
+            Err(AuthError::InvalidApiKey)
+        }
+    }
+
+    impl<const A: u8> ActionPolicy<A> {
+        fn authenticate_tenant_token(
+            auth: &AuthController,
+            token: &str,
+        ) -> Result<TenantTokenOutcome, AuthError> {
+            // Only search action can be accessed by a tenant token.
+            if A != actions::SEARCH {
+                return Ok(TenantTokenOutcome::NotATenantToken);
+            }
+
+            let uid = extract_key_id(token)?;
+
+            // Check if tenant token is valid.
+            let key = if let Some(key) = auth.generate_key(uid) {
+                key
+            } else {
+                return Err(AuthError::InvalidTenantToken);
+            };
+
+            let data = decode::<Claims>(
+                token,
+                &DecodingKey::from_secret(key.as_bytes()),
+                &tenant_token_validation(),
+            )?;
+
+            // Check if token is expired.
+            if let Some(exp) = data.claims.exp {
+                let now = OffsetDateTime::now_utc().unix_timestamp();
+                if now > exp {
+                    return Err(AuthError::ExpiredTenantToken { exp, now });
+                }
+            }
+
+            Ok(TenantTokenOutcome::Valid(uid, data.claims.search_rules))
+        }
+    }
+
+    #[derive(Debug, Serialize, Deserialize)]
+    #[serde(rename_all = "camelCase")]
+    struct Claims {
+        search_rules: SearchRules,
+        exp: Option<i64>,
+        api_key_uid: Uuid,
+    }
+}

crates/meilisearch/src/extractors/mod.rs

@@ -0,0 +1,4 @@
+pub mod payload;
+#[macro_use]
+pub mod authentication;
+pub mod sequential_extractor;

crates/meilisearch/src/extractors/payload.rs

@@ -0,0 +1,110 @@
+use std::pin::Pin;
+use std::task::{Context, Poll};
+
+use actix_http::encoding::Decoder as Decompress;
+use actix_web::{dev, web, FromRequest, HttpRequest};
+use futures::future::{ready, Ready};
+use futures::Stream;
+
+use crate::error::MeilisearchHttpError;
+
+pub struct Payload {
+    payload: Decompress<dev::Payload>,
+    limit: usize,
+    remaining: usize,
+}
+
+pub struct PayloadConfig {
+    limit: usize,
+}
+
+impl PayloadConfig {
+    pub fn new(limit: usize) -> Self {
+        Self { limit }
+    }
+}
+
+impl Default for PayloadConfig {
+    fn default() -> Self {
+        Self { limit: 256 * 1024 }
+    }
+}
+
+impl FromRequest for Payload {
+    type Error = MeilisearchHttpError;
+
+    type Future = Ready<Result<Payload, Self::Error>>;
+
+    #[inline]
+    fn from_request(req: &HttpRequest, payload: &mut dev::Payload) -> Self::Future {
+        let limit = req
+            .app_data::<PayloadConfig>()
+            .map(|c| c.limit)
+            .unwrap_or(PayloadConfig::default().limit);
+        ready(Ok(Payload {
+            payload: Decompress::from_headers(payload.take(), req.headers()),
+            limit,
+            remaining: limit,
+        }))
+    }
+}
+
+impl Stream for Payload {
+    type Item = Result<web::Bytes, MeilisearchHttpError>;
+
+    #[inline]
+    fn poll_next(mut self: Pin<&mut Self>, cx: &mut Context<'_>) -> Poll<Option<Self::Item>> {
+        match Pin::new(&mut self.payload).poll_next(cx) {
+            Poll::Ready(Some(result)) => match result {
+                Ok(bytes) => match self.remaining.checked_sub(bytes.len()) {
+                    Some(new_limit) => {
+                        self.remaining = new_limit;
+                        Poll::Ready(Some(Ok(bytes)))
+                    }
+                    None => {
+                        Poll::Ready(Some(Err(MeilisearchHttpError::PayloadTooLarge(self.limit))))
+                    }
+                },
+                x => Poll::Ready(Some(x.map_err(MeilisearchHttpError::from))),
+            },
+            otherwise => otherwise.map(|o| o.map(|o| o.map_err(MeilisearchHttpError::from))),
+        }
+    }
+}
+
+#[cfg(test)]
+mod tests {
+    use actix_http::encoding::Decoder as Decompress;
+    use actix_http::BoxedPayloadStream;
+    use bytes::Bytes;
+    use futures_util::StreamExt;
+    use meili_snap::snapshot;
+
+    use super::*;
+
+    #[actix_rt::test]
+    async fn payload_to_large() {
+        let stream = futures::stream::iter(vec![
+            Ok(Bytes::from("1")),
+            Ok(Bytes::from("2")),
+            Ok(Bytes::from("3")),
+            Ok(Bytes::from("4")),
+        ]);
+        let boxed_stream: BoxedPayloadStream = Box::pin(stream);
+        let actix_payload = dev::Payload::from(boxed_stream);
+
+        let payload = Payload {
+            limit: 3,
+            remaining: 3,
+            payload: Decompress::new(actix_payload, actix_http::ContentEncoding::Identity),
+        };
+
+        let mut enumerated_payload_stream = payload.enumerate();
+
+        while let Some((idx, chunk)) = enumerated_payload_stream.next().await {
+            if idx == 3 {
+                snapshot!(chunk.unwrap_err(), @"The provided payload reached the size limit. The maximum accepted payload size is 3 B.");
+            }
+        }
+    }
+}

crates/meilisearch/src/extractors/sequential_extractor.rs

@@ -0,0 +1,152 @@
+#![allow(non_snake_case)]
+use std::future::Future;
+use std::pin::Pin;
+use std::task::Poll;
+
+use actix_web::dev::Payload;
+use actix_web::{FromRequest, Handler, HttpRequest};
+use pin_project_lite::pin_project;
+
+/// `SeqHandler` is an actix `Handler` that enforces that extractors errors are returned in the
+/// same order as they are defined in the wrapped handler. This is needed because, by default, actix
+/// resolves the extractors concurrently, whereas we always need the authentication extractor to
+/// throw first.
+#[derive(Clone)]
+pub struct SeqHandler<H>(pub H);
+
+pub struct SeqFromRequest<T>(T);
+
+/// This macro implements `FromRequest` for arbitrary arity handler, except for one, which is
+/// useless anyway.
+macro_rules! gen_seq {
+    ($ty:ident; $($T:ident)+) => {
+        pin_project! {
+            pub struct $ty<$($T: FromRequest), +> {
+                $(
+                #[pin]
+                $T: ExtractFuture<$T::Future, $T, $T::Error>,
+                )+
+            }
+        }
+
+        impl<$($T: FromRequest), +> Future for $ty<$($T),+> {
+            type Output = Result<SeqFromRequest<($($T),+)>, actix_web::Error>;
+
+            fn poll(self: Pin<&mut Self>, cx: &mut std::task::Context<'_>) -> Poll<Self::Output> {
+                let mut this = self.project();
+
+                let mut count_fut = 0;
+                let mut count_finished = 0;
+
+                $(
+                count_fut += 1;
+                match this.$T.as_mut().project() {
+                    ExtractProj::Future { fut } => match fut.poll(cx) {
+                        Poll::Ready(Ok(output)) => {
+                            count_finished += 1;
+                            let _ = this
+                                .$T
+                                .as_mut()
+                                .project_replace(ExtractFuture::Done { output });
+                        }
+                        Poll::Ready(Err(error)) => {
+                            count_finished += 1;
+                            let _ = this
+                                .$T
+                                .as_mut()
+                                .project_replace(ExtractFuture::Error { error });
+                        }
+                        Poll::Pending => (),
+                    },
+                    ExtractProj::Done { .. } => count_finished += 1,
+                    ExtractProj::Error { .. } => {
+                        // short circuit if all previous are finished and we had an error.
+                        if count_finished == count_fut {
+                            match this.$T.project_replace(ExtractFuture::Empty) {
+                                ExtractReplaceProj::Error { error } => {
+                                    return Poll::Ready(Err(error.into()))
+                                }
+                                _ => unreachable!("Invalid future state"),
+                            }
+                        } else {
+                            count_finished += 1;
+                        }
+                    }
+                    ExtractProj::Empty => unreachable!("From request polled after being finished. {}", stringify!($T)),
+                }
+                )+
+
+                if count_fut == count_finished {
+                    let result = (
+                        $(
+                            match this.$T.project_replace(ExtractFuture::Empty) {
+                                ExtractReplaceProj::Done { output } => output,
+                                ExtractReplaceProj::Error { error } => return Poll::Ready(Err(error.into())),
+                                _ => unreachable!("Invalid future state"),
+                            },
+                        )+
+                    );
+
+                    Poll::Ready(Ok(SeqFromRequest(result)))
+                } else {
+                    Poll::Pending
+                }
+            }
+        }
+
+        impl<$($T: FromRequest,)+> FromRequest for SeqFromRequest<($($T,)+)> {
+            type Error = actix_web::Error;
+
+            type Future = $ty<$($T),+>;
+
+            fn from_request(req: &HttpRequest, payload: &mut Payload) -> Self::Future {
+                $ty {
+                $(
+                    $T: ExtractFuture::Future {
+                        fut: $T::from_request(req, payload),
+                    },
+                )+
+                }
+            }
+        }
+
+    impl<Han, $($T: FromRequest),+> Handler<SeqFromRequest<($($T),+)>> for SeqHandler<Han>
+    where
+        Han: Handler<($($T),+)>,
+    {
+        type Output = Han::Output;
+        type Future = Han::Future;
+
+        fn call(&self, args: SeqFromRequest<($($T),+)>) -> Self::Future {
+            self.0.call(args.0)
+        }
+    }
+    };
+}
+
+// Not working for a single argument, but then, it is not really necessary.
+// gen_seq! { SeqFromRequestFut1; A }
+gen_seq! { SeqFromRequestFut2; A B }
+gen_seq! { SeqFromRequestFut3; A B C }
+gen_seq! { SeqFromRequestFut4; A B C D }
+gen_seq! { SeqFromRequestFut5; A B C D E }
+gen_seq! { SeqFromRequestFut6; A B C D E F }
+gen_seq! { SeqFromRequestFut7; A B C D E F G }
+
+pin_project! {
+    #[project = ExtractProj]
+    #[project_replace = ExtractReplaceProj]
+    enum ExtractFuture<Fut, Res, Err> {
+        Future {
+            #[pin]
+            fut: Fut,
+        },
+        Done {
+            output: Res,
+        },
+        Error {
+            error: Err,
+        },
+        Empty,
+    }
+}

crates/meilisearch/src/lib.rs

@@ -0,0 +1,545 @@
+#![allow(rustdoc::private_intra_doc_links)]
+#[macro_use]
+pub mod error;
+pub mod analytics;
+#[macro_use]
+pub mod extractors;
+pub mod metrics;
+pub mod middleware;
+pub mod option;
+pub mod routes;
+pub mod search;
+pub mod search_queue;
+
+use std::fs::File;
+use std::io::{BufReader, BufWriter};
+use std::path::Path;
+use std::str::FromStr;
+use std::sync::Arc;
+use std::thread;
+use std::time::Duration;
+
+use actix_cors::Cors;
+use actix_http::body::MessageBody;
+use actix_web::dev::{ServiceFactory, ServiceResponse};
+use actix_web::error::JsonPayloadError;
+use actix_web::http::header::{CONTENT_TYPE, USER_AGENT};
+use actix_web::web::Data;
+use actix_web::{web, HttpRequest};
+use analytics::Analytics;
+use anyhow::bail;
+use error::PayloadError;
+use extractors::payload::PayloadConfig;
+use index_scheduler::{IndexScheduler, IndexSchedulerOptions};
+use meilisearch_auth::AuthController;
+use meilisearch_types::milli::documents::{DocumentsBatchBuilder, DocumentsBatchReader};
+use meilisearch_types::milli::update::{IndexDocumentsConfig, IndexDocumentsMethod};
+use meilisearch_types::settings::apply_settings_to_builder;
+use meilisearch_types::tasks::KindWithContent;
+use meilisearch_types::versioning::{check_version_file, create_current_version_file};
+use meilisearch_types::{compression, milli, VERSION_FILE_NAME};
+pub use option::Opt;
+use option::ScheduleSnapshot;
+use search_queue::SearchQueue;
+use tracing::{error, info_span};
+use tracing_subscriber::filter::Targets;
+
+use crate::error::MeilisearchHttpError;
+
+/// Default number of simultaneously opened indexes.
+///
+/// This value is used when dynamic computation of how many indexes can be opened at once was skipped (e.g., in tests).
+///
+/// Lower for Windows that dedicates a smaller virtual address space to processes.
+///
+/// The value was chosen this way:
+///
+/// - Windows provides a small virtual address space of about 10TiB to processes.
+/// - The chosen value allows for indexes to use the default map size of 2TiB safely.
+#[cfg(windows)]
+const DEFAULT_INDEX_COUNT: usize = 4;
+
+/// Default number of simultaneously opened indexes.
+///
+/// This value is used when dynamic computation of how many indexes can be opened at once was skipped (e.g., in tests).
+///
+/// The higher, the better for avoiding reopening indexes.
+///
+/// The value was chosen this way:
+///
+/// - Opening an index consumes a file descriptor.
+/// - The default on many unices is about 256 file descriptors for a process.
+/// - 100 is a little bit less than half this value.
+/// - The chosen value allows for indexes to use the default map size of 2TiB safely.
+#[cfg(not(windows))]
+const DEFAULT_INDEX_COUNT: usize = 20;
+
+/// Check if a db is empty. It does not provide any information on the
+/// validity of the data in it.
+/// We consider a database as non empty when it's a non empty directory.
+fn is_empty_db(db_path: impl AsRef<Path>) -> bool {
+    let db_path = db_path.as_ref();
+
+    if !db_path.exists() {
+        true
+    // if we encounter an error or if the db is a file we consider the db non empty
+    } else if let Ok(dir) = db_path.read_dir() {
+        dir.count() == 0
+    } else {
+        true
+    }
+}
+
+/// The handle used to update the logs at runtime. Must be accessible from the `main.rs` and the `route/logs.rs`.
+pub type LogRouteHandle =
+    tracing_subscriber::reload::Handle<LogRouteType, tracing_subscriber::Registry>;
+
+pub type LogRouteType = tracing_subscriber::filter::Filtered<
+    Option<Box<dyn tracing_subscriber::Layer<tracing_subscriber::Registry> + Send + Sync>>,
+    Targets,
+    tracing_subscriber::Registry,
+>;
+
+pub type SubscriberForSecondLayer = tracing_subscriber::layer::Layered<
+    tracing_subscriber::reload::Layer<LogRouteType, tracing_subscriber::Registry>,
+    tracing_subscriber::Registry,
+>;
+
+pub type LogStderrHandle =
+    tracing_subscriber::reload::Handle<LogStderrType, SubscriberForSecondLayer>;
+
+pub type LogStderrType = tracing_subscriber::filter::Filtered<
+    Box<dyn tracing_subscriber::Layer<SubscriberForSecondLayer> + Send + Sync>,
+    Targets,
+    SubscriberForSecondLayer,
+>;
+
+pub fn create_app(
+    index_scheduler: Data<IndexScheduler>,
+    auth_controller: Data<AuthController>,
+    search_queue: Data<SearchQueue>,
+    opt: Opt,
+    logs: (LogRouteHandle, LogStderrHandle),
+    analytics: Data<Analytics>,
+    enable_dashboard: bool,
+) -> actix_web::App<
+    impl ServiceFactory<
+        actix_web::dev::ServiceRequest,
+        Config = (),
+        Response = ServiceResponse<impl MessageBody>,
+        Error = actix_web::Error,
+        InitError = (),
+    >,
+> {
+    let app = actix_web::App::new()
+        .configure(|s| {
+            configure_data(
+                s,
+                index_scheduler.clone(),
+                auth_controller.clone(),
+                search_queue.clone(),
+                &opt,
+                logs,
+                analytics.clone(),
+            )
+        })
+        .configure(routes::configure)
+        .configure(|s| dashboard(s, enable_dashboard));
+
+    let app = app.wrap(middleware::RouteMetrics);
+    app.wrap(
+        Cors::default()
+            .send_wildcard()
+            .allow_any_header()
+            .allow_any_origin()
+            .allow_any_method()
+            .max_age(86_400), // 24h
+    )
+    .wrap(tracing_actix_web::TracingLogger::<AwebTracingLogger>::new())
+    .wrap(actix_web::middleware::Compress::default())
+    .wrap(actix_web::middleware::NormalizePath::new(actix_web::middleware::TrailingSlash::Trim))
+}
+
+struct AwebTracingLogger;
+
+impl tracing_actix_web::RootSpanBuilder for AwebTracingLogger {
+    fn on_request_start(request: &actix_web::dev::ServiceRequest) -> tracing::Span {
+        use tracing::field::Empty;
+
+        let conn_info = request.connection_info();
+        let headers = request.headers();
+        let user_agent = headers
+            .get(USER_AGENT)
+            .map(|value| String::from_utf8_lossy(value.as_bytes()).into_owned())
+            .unwrap_or_default();
+        info_span!("HTTP request", method = %request.method(), host = conn_info.host(), route = %request.path(), query_parameters = %request.query_string(), %user_agent, status_code = Empty, error = Empty)
+    }
+
+    fn on_request_end<B: MessageBody>(
+        span: tracing::Span,
+        outcome: &Result<ServiceResponse<B>, actix_web::Error>,
+    ) {
+        match &outcome {
+            Ok(response) => {
+                let code: i32 = response.response().status().as_u16().into();
+                span.record("status_code", code);
+
+                if let Some(error) = response.response().error() {
+                    // use the status code already constructed for the outgoing HTTP response
+                    span.record("error", &tracing::field::display(error.as_response_error()));
+                }
+            }
+            Err(error) => {
+                let code: i32 = error.error_response().status().as_u16().into();
+                span.record("status_code", code);
+                span.record("error", &tracing::field::display(error.as_response_error()));
+            }
+        };
+    }
+}
+
+enum OnFailure {
+    RemoveDb,
+    KeepDb,
+}
+
+pub fn setup_meilisearch(opt: &Opt) -> anyhow::Result<(Arc<IndexScheduler>, Arc<AuthController>)> {
+    let empty_db = is_empty_db(&opt.db_path);
+    let (index_scheduler, auth_controller) = if let Some(ref snapshot_path) = opt.import_snapshot {
+        let snapshot_path_exists = snapshot_path.exists();
+        // the db is empty and the snapshot exists, import it
+        if empty_db && snapshot_path_exists {
+            match compression::from_tar_gz(snapshot_path, &opt.db_path) {
+                Ok(()) => open_or_create_database_unchecked(opt, OnFailure::RemoveDb)?,
+                Err(e) => {
+                    std::fs::remove_dir_all(&opt.db_path)?;
+                    return Err(e);
+                }
+            }
+        // the db already exists and we should not ignore the snapshot => throw an error
+        } else if !empty_db && !opt.ignore_snapshot_if_db_exists {
+            bail!(
+                "database already exists at {:?}, try to delete it or rename it",
+                opt.db_path.canonicalize().unwrap_or_else(|_| opt.db_path.to_owned())
+            )
+        // the snapshot doesn't exist and we can't ignore it => throw an error
+        } else if !snapshot_path_exists && !opt.ignore_missing_snapshot {
+            bail!("snapshot doesn't exist at {}", snapshot_path.display())
+        // the snapshot and the db exist, and we can ignore the snapshot because of the ignore_snapshot_if_db_exists flag
+        } else {
+            open_or_create_database(opt, empty_db)?
+        }
+    } else if let Some(ref path) = opt.import_dump {
+        let src_path_exists = path.exists();
+        // the db is empty and the dump exists, import it
+        if empty_db && src_path_exists {
+            let (mut index_scheduler, mut auth_controller) =
+                open_or_create_database_unchecked(opt, OnFailure::RemoveDb)?;
+            match import_dump(&opt.db_path, path, &mut index_scheduler, &mut auth_controller) {
+                Ok(()) => (index_scheduler, auth_controller),
+                Err(e) => {
+                    std::fs::remove_dir_all(&opt.db_path)?;
+                    return Err(e);
+                }
+            }
+        // the db already exists and we should not ignore the dump option => throw an error
+        } else if !empty_db && !opt.ignore_dump_if_db_exists {
+            bail!(
+                "database already exists at {:?}, try to delete it or rename it",
+                opt.db_path.canonicalize().unwrap_or_else(|_| opt.db_path.to_owned())
+            )
+        // the dump doesn't exist and we can't ignore it => throw an error
+        } else if !src_path_exists && !opt.ignore_missing_dump {
+            bail!("dump doesn't exist at {:?}", path)
+        // the dump and the db exist and we can ignore the dump because of the ignore_dump_if_db_exists flag
+        // or, the dump is missing but we can ignore that because of the ignore_missing_dump flag
+        } else {
+            open_or_create_database(opt, empty_db)?
+        }
+    } else {
+        open_or_create_database(opt, empty_db)?
+    };
+
+    // We create a loop in a thread that registers snapshotCreation tasks
+    let index_scheduler = Arc::new(index_scheduler);
+    let auth_controller = Arc::new(auth_controller);
+    if let ScheduleSnapshot::Enabled(snapshot_delay) = opt.schedule_snapshot {
+        let snapshot_delay = Duration::from_secs(snapshot_delay);
+        let index_scheduler = index_scheduler.clone();
+        thread::Builder::new()
+            .name(String::from("register-snapshot-tasks"))
+            .spawn(move || loop {
+                thread::sleep(snapshot_delay);
+                if let Err(e) =
+                    index_scheduler.register(KindWithContent::SnapshotCreation, None, false)
+                {
+                    error!("Error while registering snapshot: {}", e);
+                }
+            })
+            .unwrap();
+    }
+
+    Ok((index_scheduler, auth_controller))
+}
+
+/// Try to start the IndexScheduler and AuthController without checking the VERSION file or anything.
+fn open_or_create_database_unchecked(
+    opt: &Opt,
+    on_failure: OnFailure,
+) -> anyhow::Result<(IndexScheduler, AuthController)> {
+    // we don't want to create anything in the data.ms yet, thus we
+    // wrap our two builders in a closure that'll be executed later.
+    let auth_controller = AuthController::new(&opt.db_path, &opt.master_key);
+    let instance_features = opt.to_instance_features();
+    let index_scheduler_builder = || -> anyhow::Result<_> {
+        Ok(IndexScheduler::new(IndexSchedulerOptions {
+            version_file_path: opt.db_path.join(VERSION_FILE_NAME),
+            auth_path: opt.db_path.join("auth"),
+            tasks_path: opt.db_path.join("tasks"),
+            update_file_path: opt.db_path.join("update_files"),
+            indexes_path: opt.db_path.join("indexes"),
+            snapshots_path: opt.snapshot_dir.clone(),
+            dumps_path: opt.dump_dir.clone(),
+            webhook_url: opt.task_webhook_url.as_ref().map(|url| url.to_string()),
+            webhook_authorization_header: opt.task_webhook_authorization_header.clone(),
+            task_db_size: opt.max_task_db_size.as_u64() as usize,
+            index_base_map_size: opt.max_index_size.as_u64() as usize,
+            enable_mdb_writemap: opt.experimental_reduce_indexing_memory_usage,
+            indexer_config: (&opt.indexer_options).try_into()?,
+            autobatching_enabled: true,
+            cleanup_enabled: !opt.experimental_replication_parameters,
+            max_number_of_tasks: 1_000_000,
+            max_number_of_batched_tasks: opt.experimental_max_number_of_batched_tasks,
+            index_growth_amount: byte_unit::Byte::from_str("10GiB").unwrap().as_u64() as usize,
+            index_count: DEFAULT_INDEX_COUNT,
+            instance_features,
+        })?)
+    };
+
+    match (
+        index_scheduler_builder(),
+        auth_controller.map_err(anyhow::Error::from),
+        create_current_version_file(&opt.db_path).map_err(anyhow::Error::from),
+    ) {
+        (Ok(i), Ok(a), Ok(())) => Ok((i, a)),
+        (Err(e), _, _) | (_, Err(e), _) | (_, _, Err(e)) => {
+            if matches!(on_failure, OnFailure::RemoveDb) {
+                std::fs::remove_dir_all(&opt.db_path)?;
+            }
+            Err(e)
+        }
+    }
+}
+
+/// Ensure you're in a valid state and open the IndexScheduler + AuthController for you.
+fn open_or_create_database(
+    opt: &Opt,
+    empty_db: bool,
+) -> anyhow::Result<(IndexScheduler, AuthController)> {
+    if !empty_db {
+        check_version_file(&opt.db_path)?;
+    }
+
+    open_or_create_database_unchecked(opt, OnFailure::KeepDb)
+}
+
+fn import_dump(
+    db_path: &Path,
+    dump_path: &Path,
+    index_scheduler: &mut IndexScheduler,
+    auth: &mut AuthController,
+) -> Result<(), anyhow::Error> {
+    let reader = File::open(dump_path)?;
+    let mut dump_reader = dump::DumpReader::open(reader)?;
+
+    if let Some(date) = dump_reader.date() {
+        tracing::info!(
+            version = ?dump_reader.version(), // TODO: get the meilisearch version instead of the dump version
+            %date,
+            "Importing a dump of meilisearch"
+        );
+    } else {
+        tracing::info!(
+            version = ?dump_reader.version(), // TODO: get the meilisearch version instead of the dump version
+            "Importing a dump of meilisearch",
+        );
+    }
+
+    let instance_uid = dump_reader.instance_uid()?;
+
+    // 1. Import the instance-uid.
+    if let Some(ref instance_uid) = instance_uid {
+        // we don't want to panic if there is an error with the instance-uid.
+        let _ = std::fs::write(db_path.join("instance-uid"), instance_uid.to_string().as_bytes());
+    };
+
+    // 2. Import the `Key`s.
+    let mut keys = Vec::new();
+    auth.raw_delete_all_keys()?;
+    for key in dump_reader.keys()? {
+        let key = key?;
+        auth.raw_insert_key(key.clone())?;
+        keys.push(key);
+    }
+
+    // 3. Import the runtime features.
+    let features = dump_reader.features()?.unwrap_or_default();
+    index_scheduler.put_runtime_features(features)?;
+
+    let indexer_config = index_scheduler.indexer_config();
+
+    // /!\ The tasks must be imported AFTER importing the indexes or else the scheduler might
+    // try to process tasks while we're trying to import the indexes.
+
+    // 4. Import the indexes.
+    for index_reader in dump_reader.indexes()? {
+        let mut index_reader = index_reader?;
+        let metadata = index_reader.metadata();
+        tracing::info!("Importing index `{}`.", metadata.uid);
+
+        let date = Some((metadata.created_at, metadata.updated_at));
+        let index = index_scheduler.create_raw_index(&metadata.uid, date)?;
+
+        let mut wtxn = index.write_txn()?;
+
+        let mut builder = milli::update::Settings::new(&mut wtxn, &index, indexer_config);
+        // 4.1 Import the primary key if there is one.
+        if let Some(ref primary_key) = metadata.primary_key {
+            builder.set_primary_key(primary_key.to_string());
+        }
+
+        // 4.2 Import the settings.
+        tracing::info!("Importing the settings.");
+        let settings = index_reader.settings()?;
+        apply_settings_to_builder(&settings, &mut builder);
+        builder
+            .execute(|indexing_step| tracing::debug!("update: {:?}", indexing_step), || false)?;
+
+        // 4.3 Import the documents.
+        // 4.3.1 We need to recreate the grenad+obkv format accepted by the index.
+        tracing::info!("Importing the documents.");
+        let file = tempfile::tempfile()?;
+        let mut builder = DocumentsBatchBuilder::new(BufWriter::new(file));
+        for document in index_reader.documents()? {
+            builder.append_json_object(&document?)?;
+        }
+
+        // This flush the content of the batch builder.
+        let file = builder.into_inner()?.into_inner()?;
+
+        // 4.3.2 We feed it to the milli index.
+        let reader = BufReader::new(file);
+        let reader = DocumentsBatchReader::from_reader(reader)?;
+
+        let embedder_configs = index.embedding_configs(&wtxn)?;
+        let embedders = index_scheduler.embedders(embedder_configs)?;
+
+        let builder = milli::update::IndexDocuments::new(
+            &mut wtxn,
+            &index,
+            indexer_config,
+            IndexDocumentsConfig {
+                update_method: IndexDocumentsMethod::ReplaceDocuments,
+                ..Default::default()
+            },
+            |indexing_step| tracing::trace!("update: {:?}", indexing_step),
+            || false,
+        )?;
+
+        let builder = builder.with_embedders(embedders);
+
+        let (builder, user_result) = builder.add_documents(reader)?;
+        let user_result = user_result?;
+        tracing::info!(documents_found = user_result, "{} documents found.", user_result);
+        builder.execute()?;
+        wtxn.commit()?;
+        tracing::info!("All documents successfully imported.");
+    }
+
+    let mut index_scheduler_dump = index_scheduler.register_dumped_task()?;
+
+    // 5. Import the tasks.
+    for ret in dump_reader.tasks()? {
+        let (task, file) = ret?;
+        index_scheduler_dump.register_dumped_task(task, file)?;
+    }
+    Ok(index_scheduler_dump.finish()?)
+}
+
+pub fn configure_data(
+    config: &mut web::ServiceConfig,
+    index_scheduler: Data<IndexScheduler>,
+    auth: Data<AuthController>,
+    search_queue: Data<SearchQueue>,
+    opt: &Opt,
+    (logs_route, logs_stderr): (LogRouteHandle, LogStderrHandle),
+    analytics: Data<Analytics>,
+) {
+    let http_payload_size_limit = opt.http_payload_size_limit.as_u64() as usize;
+    config
+        .app_data(index_scheduler)
+        .app_data(auth)
+        .app_data(search_queue)
+        .app_data(analytics)
+        .app_data(web::Data::new(logs_route))
+        .app_data(web::Data::new(logs_stderr))
+        .app_data(web::Data::new(opt.clone()))
+        .app_data(
+            web::JsonConfig::default()
+                .limit(http_payload_size_limit)
+                .content_type(|mime| mime == mime::APPLICATION_JSON)
+                .error_handler(|err, req: &HttpRequest| match err {
+                    JsonPayloadError::ContentType => match req.headers().get(CONTENT_TYPE) {
+                        Some(content_type) => MeilisearchHttpError::InvalidContentType(
+                            content_type.to_str().unwrap_or("unknown").to_string(),
+                            vec![mime::APPLICATION_JSON.to_string()],
+                        )
+                        .into(),
+                        None => MeilisearchHttpError::MissingContentType(vec![
+                            mime::APPLICATION_JSON.to_string(),
+                        ])
+                        .into(),
+                    },
+                    err => PayloadError::from(err).into(),
+                }),
+        )
+        .app_data(PayloadConfig::new(http_payload_size_limit))
+        .app_data(
+            web::QueryConfig::default().error_handler(|err, _req| PayloadError::from(err).into()),
+        );
+}
+
+#[cfg(feature = "mini-dashboard")]
+pub fn dashboard(config: &mut web::ServiceConfig, enable_frontend: bool) {
+    use actix_web::HttpResponse;
+    use static_files::Resource;
+
+    mod generated {
+        include!(concat!(env!("OUT_DIR"), "/generated.rs"));
+    }
+
+    if enable_frontend {
+        let generated = generated::generate();
+        // Generate routes for mini-dashboard assets
+        for (path, resource) in generated.into_iter() {
+            let Resource { mime_type, data, .. } = resource;
+            // Redirect index.html to /
+            if path == "index.html" {
+                config.service(web::resource("/").route(web::get().to(move || async move {
+                    HttpResponse::Ok().content_type(mime_type).body(data)
+                })));
+            } else {
+                config.service(web::resource(path).route(web::get().to(move || async move {
+                    HttpResponse::Ok().content_type(mime_type).body(data)
+                })));
+            }
+        }
+    } else {
+        config.service(web::resource("/").route(web::get().to(routes::running)));
+    }
+}
+
+#[cfg(not(feature = "mini-dashboard"))]
+pub fn dashboard(config: &mut web::ServiceConfig, _enable_frontend: bool) {
+    config.service(web::resource("/").route(web::get().to(routes::running)));
+}

crates/meilisearch/src/main.rs

@@ -0,0 +1,331 @@
+use std::env;
+use std::io::{stderr, LineWriter, Write};
+use std::num::NonZeroUsize;
+use std::path::PathBuf;
+use std::str::FromStr;
+use std::sync::Arc;
+use std::thread::available_parallelism;
+use std::time::Duration;
+
+use actix_web::http::KeepAlive;
+use actix_web::web::Data;
+use actix_web::HttpServer;
+use index_scheduler::IndexScheduler;
+use is_terminal::IsTerminal;
+use meilisearch::analytics::Analytics;
+use meilisearch::option::LogMode;
+use meilisearch::search_queue::SearchQueue;
+use meilisearch::{
+    analytics, create_app, setup_meilisearch, LogRouteHandle, LogRouteType, LogStderrHandle,
+    LogStderrType, Opt, SubscriberForSecondLayer,
+};
+use meilisearch_auth::{generate_master_key, AuthController, MASTER_KEY_MIN_SIZE};
+use mimalloc::MiMalloc;
+use termcolor::{Color, ColorChoice, ColorSpec, StandardStream, WriteColor};
+use tracing::level_filters::LevelFilter;
+use tracing_subscriber::layer::SubscriberExt as _;
+use tracing_subscriber::Layer;
+
+#[global_allocator]
+static ALLOC: MiMalloc = MiMalloc;
+
+fn default_log_route_layer() -> LogRouteType {
+    None.with_filter(tracing_subscriber::filter::Targets::new().with_target("", LevelFilter::OFF))
+}
+
+fn default_log_stderr_layer(opt: &Opt) -> LogStderrType {
+    let layer = tracing_subscriber::fmt::layer()
+        .with_writer(|| LineWriter::new(std::io::stderr()))
+        .with_span_events(tracing_subscriber::fmt::format::FmtSpan::CLOSE);
+
+    let layer = match opt.experimental_logs_mode {
+        LogMode::Human => Box::new(layer)
+            as Box<dyn tracing_subscriber::Layer<SubscriberForSecondLayer> + Send + Sync>,
+        LogMode::Json => Box::new(layer.json())
+            as Box<dyn tracing_subscriber::Layer<SubscriberForSecondLayer> + Send + Sync>,
+    };
+
+    layer.with_filter(
+        tracing_subscriber::filter::Targets::new()
+            .with_target("", LevelFilter::from_str(&opt.log_level.to_string()).unwrap()),
+    )
+}
+
+/// does all the setup before meilisearch is launched
+fn setup(opt: &Opt) -> anyhow::Result<(LogRouteHandle, LogStderrHandle)> {
+    let (route_layer, route_layer_handle) =
+        tracing_subscriber::reload::Layer::new(default_log_route_layer());
+    let route_layer: tracing_subscriber::reload::Layer<_, _> = route_layer;
+
+    let (stderr_layer, stderr_layer_handle) =
+        tracing_subscriber::reload::Layer::new(default_log_stderr_layer(opt));
+    let route_layer: tracing_subscriber::reload::Layer<_, _> = route_layer;
+
+    let subscriber = tracing_subscriber::registry().with(route_layer).with(stderr_layer);
+
+    // set the subscriber as the default for the application
+    tracing::subscriber::set_global_default(subscriber).unwrap();
+
+    Ok((route_layer_handle, stderr_layer_handle))
+}
+
+fn on_panic(info: &std::panic::PanicInfo) {
+    let info = info.to_string().replace('\n', " ");
+    tracing::error!(%info);
+}
+
+#[actix_web::main]
+async fn main() -> anyhow::Result<()> {
+    try_main().await.inspect_err(|error| {
+        tracing::error!(%error);
+        let mut current = error.source();
+        let mut depth = 0;
+        while let Some(source) = current {
+            tracing::info!(%source, depth, "Error caused by");
+            current = source.source();
+            depth += 1;
+        }
+    })
+}
+
+async fn try_main() -> anyhow::Result<()> {
+    let (opt, config_read_from) = Opt::try_build()?;
+
+    std::panic::set_hook(Box::new(on_panic));
+
+    anyhow::ensure!(
+        !(cfg!(windows) && opt.experimental_reduce_indexing_memory_usage),
+        "The `experimental-reduce-indexing-memory-usage` flag is not supported on Windows"
+    );
+
+    let log_handle = setup(&opt)?;
+
+    match (opt.env.as_ref(), &opt.master_key) {
+        ("production", Some(master_key)) if master_key.len() < MASTER_KEY_MIN_SIZE => {
+            anyhow::bail!(
+                "The master key must be at least {MASTER_KEY_MIN_SIZE} bytes in a production environment. The provided key is only {} bytes.
+
+{}",
+                master_key.len(),
+                generated_master_key_message(),
+            )
+        }
+        ("production", None) => {
+            anyhow::bail!(
+                "You must provide a master key to secure your instance in a production environment. It can be specified via the MEILI_MASTER_KEY environment variable or the --master-key launch option.
+
+{}",
+                generated_master_key_message()
+            )
+        }
+        // No error; continue
+        _ => (),
+    }
+
+    let (index_scheduler, auth_controller) = setup_meilisearch(&opt)?;
+
+    let analytics =
+        analytics::Analytics::new(&opt, index_scheduler.clone(), auth_controller.clone()).await;
+
+    print_launch_resume(&opt, analytics.clone(), config_read_from);
+
+    run_http(index_scheduler, auth_controller, opt, log_handle, Arc::new(analytics)).await?;
+
+    Ok(())
+}
+
+async fn run_http(
+    index_scheduler: Arc<IndexScheduler>,
+    auth_controller: Arc<AuthController>,
+    opt: Opt,
+    logs: (LogRouteHandle, LogStderrHandle),
+    analytics: Arc<Analytics>,
+) -> anyhow::Result<()> {
+    let enable_dashboard = &opt.env == "development";
+    let opt_clone = opt.clone();
+    let index_scheduler = Data::from(index_scheduler);
+    let auth_controller = Data::from(auth_controller);
+    let analytics = Data::from(analytics);
+    let search_queue = SearchQueue::new(
+        opt.experimental_search_queue_size,
+        available_parallelism()
+            .unwrap_or(NonZeroUsize::new(2).unwrap())
+            .checked_mul(opt.experimental_nb_searches_per_core)
+            .unwrap_or(NonZeroUsize::MAX),
+    )
+    .with_time_to_abort(Duration::from_secs(
+        usize::from(opt.experimental_drop_search_after) as u64
+    ));
+    let search_queue = Data::new(search_queue);
+
+    let http_server = HttpServer::new(move || {
+        create_app(
+            index_scheduler.clone(),
+            auth_controller.clone(),
+            search_queue.clone(),
+            opt.clone(),
+            logs.clone(),
+            analytics.clone(),
+            enable_dashboard,
+        )
+    })
+    // Disable signals allows the server to terminate immediately when a user enter CTRL-C
+    .disable_signals()
+    .keep_alive(KeepAlive::Os);
+
+    if let Some(config) = opt_clone.get_ssl_config()? {
+        http_server.bind_rustls_0_23(opt_clone.http_addr, config)?.run().await?;
+    } else {
+        http_server.bind(&opt_clone.http_addr)?.run().await?;
+    }
+    Ok(())
+}
+
+pub fn print_launch_resume(opt: &Opt, analytics: Analytics, config_read_from: Option<PathBuf>) {
+    let build_info = build_info::BuildInfo::from_build();
+
+    let protocol =
+        if opt.ssl_cert_path.is_some() && opt.ssl_key_path.is_some() { "https" } else { "http" };
+    let ascii_name = r#"
+888b     d888          d8b 888 d8b                                            888
+8888b   d8888          Y8P 888 Y8P                                            888
+88888b.d88888              888                                                888
+888Y88888P888  .d88b.  888 888 888 .d8888b   .d88b.   8888b.  888d888 .d8888b 88888b.
+888 Y888P 888 d8P  Y8b 888 888 888 88K      d8P  Y8b     "88b 888P"  d88P"    888 "88b
+888  Y8P  888 88888888 888 888 888 "Y8888b. 88888888 .d888888 888    888      888  888
+888   "   888 Y8b.     888 888 888      X88 Y8b.     888  888 888    Y88b.    888  888
+888       888  "Y8888  888 888 888  88888P'  "Y8888  "Y888888 888     "Y8888P 888  888
+"#;
+
+    eprintln!("{}", ascii_name);
+
+    eprintln!(
+        "Config file path:\t{:?}",
+        config_read_from
+            .map(|config_file_path| config_file_path.display().to_string())
+            .unwrap_or_else(|| "none".to_string())
+    );
+    eprintln!("Database path:\t\t{:?}", opt.db_path);
+    eprintln!("Server listening on:\t\"{}://{}\"", protocol, opt.http_addr);
+    eprintln!("Environment:\t\t{:?}", opt.env);
+    eprintln!("Commit SHA:\t\t{:?}", build_info.commit_sha1.unwrap_or("unknown"));
+    eprintln!(
+        "Commit date:\t\t{:?}",
+        build_info
+            .commit_timestamp
+            .and_then(|commit_timestamp| commit_timestamp
+                .format(&time::format_description::well_known::Rfc3339)
+                .ok())
+            .unwrap_or("unknown".into())
+    );
+    eprintln!("Package version:\t{:?}", env!("CARGO_PKG_VERSION").to_string());
+    if let Some(prototype) = build_info.describe.and_then(|describe| describe.as_prototype()) {
+        eprintln!("Prototype:\t\t{:?}", prototype);
+    }
+
+    {
+        if !opt.no_analytics {
+            eprintln!(
+                "
+Thank you for using Meilisearch!
+
+\nWe collect anonymized analytics to improve our product and your experience. To learn more, including how to turn off analytics, visit our dedicated documentation page: https://www.meilisearch.com/docs/learn/what_is_meilisearch/telemetry
+
+Anonymous telemetry:\t\"Enabled\""
+            );
+        } else {
+            eprintln!("Anonymous telemetry:\t\"Disabled\"");
+        }
+    }
+
+    if let Some(instance_uid) = analytics.instance_uid() {
+        eprintln!("Instance UID:\t\t\"{}\"", instance_uid);
+    }
+
+    eprintln!();
+
+    match (opt.env.as_ref(), &opt.master_key) {
+        ("production", Some(_)) => {
+            eprintln!("A master key has been set. Requests to Meilisearch won't be authorized unless you provide an authentication key.");
+        }
+        ("development", Some(master_key)) => {
+            eprintln!("A master key has been set. Requests to Meilisearch won't be authorized unless you provide an authentication key.");
+
+            if master_key.len() < MASTER_KEY_MIN_SIZE {
+                print_master_key_too_short_warning()
+            }
+        }
+        ("development", None) => print_missing_master_key_warning(),
+        // unreachable because Opt::try_build above would have failed already if any other value had been produced
+        _ => unreachable!(),
+    }
+
+    eprintln!();
+    eprintln!("Check out Meilisearch Cloud!\thttps://www.meilisearch.com/cloud?utm_campaign=oss&utm_source=engine&utm_medium=cli");
+    eprintln!("Documentation:\t\t\thttps://www.meilisearch.com/docs");
+    eprintln!("Source code:\t\t\thttps://github.com/meilisearch/meilisearch");
+    eprintln!("Discord:\t\t\thttps://discord.meilisearch.com");
+    eprintln!();
+}
+
+const WARNING_BG_COLOR: Option<Color> = Some(Color::Ansi256(178));
+const WARNING_FG_COLOR: Option<Color> = Some(Color::Ansi256(0));
+
+fn print_master_key_too_short_warning() {
+    let choice = if stderr().is_terminal() { ColorChoice::Auto } else { ColorChoice::Never };
+    let mut stderr = StandardStream::stderr(choice);
+    stderr
+        .set_color(
+            ColorSpec::new().set_bg(WARNING_BG_COLOR).set_fg(WARNING_FG_COLOR).set_bold(true),
+        )
+        .unwrap();
+    writeln!(stderr, "\n").unwrap();
+    writeln!(
+        stderr,
+        " Meilisearch started with a master key considered unsafe for use in a production environment.
+
+ A master key of at least {MASTER_KEY_MIN_SIZE} bytes will be required when switching to a production environment."
+    )
+    .unwrap();
+    stderr.reset().unwrap();
+    writeln!(stderr).unwrap();
+
+    eprintln!("\n{}", generated_master_key_message());
+    eprintln!(
+        "\nRestart Meilisearch with the argument above to use this new and secure master key."
+    )
+}
+
+fn print_missing_master_key_warning() {
+    let choice = if stderr().is_terminal() { ColorChoice::Auto } else { ColorChoice::Never };
+    let mut stderr = StandardStream::stderr(choice);
+    stderr
+        .set_color(
+            ColorSpec::new().set_bg(WARNING_BG_COLOR).set_fg(WARNING_FG_COLOR).set_bold(true),
+        )
+        .unwrap();
+    writeln!(stderr, "\n").unwrap();
+    writeln!(
+    stderr,
+    " No master key was found. The server will accept unidentified requests.
+
+ A master key of at least {MASTER_KEY_MIN_SIZE} bytes will be required when switching to a production environment."
+)
+.unwrap();
+    stderr.reset().unwrap();
+    writeln!(stderr).unwrap();
+
+    eprintln!("\n{}", generated_master_key_message());
+    eprintln!(
+        "\nRestart Meilisearch with the argument above to use this new and secure master key."
+    )
+}
+
+fn generated_master_key_message() -> String {
+    format!(
+        "We generated a new secure master key for you (you can safely use this token):
+
+>> --master-key {} <<",
+        generate_master_key()
+    )
+}

crates/meilisearch/src/metrics.rs

@@ -0,0 +1,52 @@
+use lazy_static::lazy_static;
+use prometheus::{
+    opts, register_histogram_vec, register_int_counter_vec, register_int_gauge,
+    register_int_gauge_vec, HistogramVec, IntCounterVec, IntGauge, IntGaugeVec,
+};
+
+lazy_static! {
+    pub static ref MEILISEARCH_HTTP_REQUESTS_TOTAL: IntCounterVec = register_int_counter_vec!(
+        opts!("meilisearch_http_requests_total", "Meilisearch HTTP requests total"),
+        &["method", "path", "status"]
+    )
+    .expect("Can't create a metric");
+    pub static ref MEILISEARCH_DEGRADED_SEARCH_REQUESTS: IntGauge = register_int_gauge!(opts!(
+        "meilisearch_degraded_search_requests",
+        "Meilisearch number of degraded search requests"
+    ))
+    .expect("Can't create a metric");
+    pub static ref MEILISEARCH_DB_SIZE_BYTES: IntGauge =
+        register_int_gauge!(opts!("meilisearch_db_size_bytes", "Meilisearch DB Size In Bytes"))
+            .expect("Can't create a metric");
+    pub static ref MEILISEARCH_USED_DB_SIZE_BYTES: IntGauge = register_int_gauge!(opts!(
+        "meilisearch_used_db_size_bytes",
+        "Meilisearch Used DB Size In Bytes"
+    ))
+    .expect("Can't create a metric");
+    pub static ref MEILISEARCH_INDEX_COUNT: IntGauge =
+        register_int_gauge!(opts!("meilisearch_index_count", "Meilisearch Index Count"))
+            .expect("Can't create a metric");
+    pub static ref MEILISEARCH_INDEX_DOCS_COUNT: IntGaugeVec = register_int_gauge_vec!(
+        opts!("meilisearch_index_docs_count", "Meilisearch Index Docs Count"),
+        &["index"]
+    )
+    .expect("Can't create a metric");
+    pub static ref MEILISEARCH_HTTP_RESPONSE_TIME_SECONDS: HistogramVec = register_histogram_vec!(
+        "meilisearch_http_response_time_seconds",
+        "Meilisearch HTTP response times",
+        &["method", "path"],
+        vec![0.005, 0.01, 0.025, 0.05, 0.075, 0.1, 0.25, 0.5, 0.75, 1.0, 2.5, 5.0, 7.5, 10.0]
+    )
+    .expect("Can't create a metric");
+    pub static ref MEILISEARCH_NB_TASKS: IntGaugeVec = register_int_gauge_vec!(
+        opts!("meilisearch_nb_tasks", "Meilisearch Number of tasks"),
+        &["kind", "value"]
+    )
+    .expect("Can't create a metric");
+    pub static ref MEILISEARCH_LAST_UPDATE: IntGauge =
+        register_int_gauge!(opts!("meilisearch_last_update", "Meilisearch Last Update"))
+            .expect("Can't create a metric");
+    pub static ref MEILISEARCH_IS_INDEXING: IntGauge =
+        register_int_gauge!(opts!("meilisearch_is_indexing", "Meilisearch Is Indexing"))
+            .expect("Can't create a metric");
+}

crates/meilisearch/src/middleware.rs

@@ -0,0 +1,89 @@
+//! Contains all the custom middleware used in meilisearch
+
+use std::future::{ready, Ready};
+
+use actix_web::dev::{self, Service, ServiceRequest, ServiceResponse, Transform};
+use actix_web::web::Data;
+use actix_web::Error;
+use futures_util::future::LocalBoxFuture;
+use index_scheduler::IndexScheduler;
+use prometheus::HistogramTimer;
+
+pub struct RouteMetrics;
+
+// Middleware factory is `Transform` trait from actix-service crate
+// `S` - type of the next service
+// `B` - type of response's body
+impl<S, B> Transform<S, ServiceRequest> for RouteMetrics
+where
+    S: Service<ServiceRequest, Response = ServiceResponse<B>, Error = Error>,
+    S::Future: 'static,
+    B: 'static,
+{
+    type Response = ServiceResponse<B>;
+    type Error = Error;
+    type InitError = ();
+    type Transform = RouteMetricsMiddleware<S>;
+    type Future = Ready<Result<Self::Transform, Self::InitError>>;
+
+    fn new_transform(&self, service: S) -> Self::Future {
+        ready(Ok(RouteMetricsMiddleware { service }))
+    }
+}
+
+pub struct RouteMetricsMiddleware<S> {
+    service: S,
+}
+
+impl<S, B> Service<ServiceRequest> for RouteMetricsMiddleware<S>
+where
+    S: Service<ServiceRequest, Response = ServiceResponse<B>, Error = Error>,
+    S::Future: 'static,
+    B: 'static,
+{
+    type Response = ServiceResponse<B>;
+    type Error = Error;
+    type Future = LocalBoxFuture<'static, Result<Self::Response, Self::Error>>;
+
+    dev::forward_ready!(service);
+
+    fn call(&self, req: ServiceRequest) -> Self::Future {
+        let mut histogram_timer: Option<HistogramTimer> = None;
+
+        // calling unwrap here is safe because index scheduler is added to app data while creating actix app.
+        // also, the tests will fail if this is not present.
+        let index_scheduler = req.app_data::<Data<IndexScheduler>>().unwrap();
+        let features = index_scheduler.features();
+
+        let request_path = req.path();
+        let request_pattern = req.match_pattern();
+        let metric_path = request_pattern.as_ref().map_or(request_path, String::as_str).to_string();
+        let request_method = req.method().to_string();
+
+        if features.check_metrics().is_ok() {
+            let is_registered_resource = req.resource_map().has_resource(request_path);
+            if is_registered_resource {
+                histogram_timer = Some(
+                    crate::metrics::MEILISEARCH_HTTP_RESPONSE_TIME_SECONDS
+                        .with_label_values(&[&request_method, &metric_path])
+                        .start_timer(),
+                );
+            }
+        };
+
+        let fut = self.service.call(req);
+
+        Box::pin(async move {
+            let res = fut.await?;
+
+            crate::metrics::MEILISEARCH_HTTP_REQUESTS_TOTAL
+                .with_label_values(&[&request_method, &metric_path, res.status().as_str()])
+                .inc();
+
+            if let Some(histogram_timer) = histogram_timer {
+                histogram_timer.observe_duration();
+            };
+            Ok(res)
+        })
+    }
+}

crates/meilisearch/src/routes/api_key.rs

@@ -0,0 +1,180 @@
+use std::str;
+
+use actix_web::web::Data;
+use actix_web::{web, HttpRequest, HttpResponse};
+use deserr::actix_web::{AwebJson, AwebQueryParameter};
+use deserr::Deserr;
+use meilisearch_auth::error::AuthControllerError;
+use meilisearch_auth::AuthController;
+use meilisearch_types::deserr::query_params::Param;
+use meilisearch_types::deserr::{DeserrJsonError, DeserrQueryParamError};
+use meilisearch_types::error::deserr_codes::*;
+use meilisearch_types::error::{Code, ResponseError};
+use meilisearch_types::keys::{CreateApiKey, Key, PatchApiKey};
+use serde::{Deserialize, Serialize};
+use time::OffsetDateTime;
+use uuid::Uuid;
+
+use super::PAGINATION_DEFAULT_LIMIT;
+use crate::extractors::authentication::policies::*;
+use crate::extractors::authentication::GuardedData;
+use crate::extractors::sequential_extractor::SeqHandler;
+use crate::routes::Pagination;
+
+pub fn configure(cfg: &mut web::ServiceConfig) {
+    cfg.service(
+        web::resource("")
+            .route(web::post().to(SeqHandler(create_api_key)))
+            .route(web::get().to(SeqHandler(list_api_keys))),
+    )
+    .service(
+        web::resource("/{key}")
+            .route(web::get().to(SeqHandler(get_api_key)))
+            .route(web::patch().to(SeqHandler(patch_api_key)))
+            .route(web::delete().to(SeqHandler(delete_api_key))),
+    );
+}
+
+pub async fn create_api_key(
+    auth_controller: GuardedData<ActionPolicy<{ actions::KEYS_CREATE }>, Data<AuthController>>,
+    body: AwebJson<CreateApiKey, DeserrJsonError>,
+    _req: HttpRequest,
+) -> Result<HttpResponse, ResponseError> {
+    let v = body.into_inner();
+    let res = tokio::task::spawn_blocking(move || -> Result<_, AuthControllerError> {
+        let key = auth_controller.create_key(v)?;
+        Ok(KeyView::from_key(key, &auth_controller))
+    })
+    .await
+    .map_err(|e| ResponseError::from_msg(e.to_string(), Code::Internal))??;
+
+    Ok(HttpResponse::Created().json(res))
+}
+
+#[derive(Deserr, Debug, Clone, Copy)]
+#[deserr(error = DeserrQueryParamError, rename_all = camelCase, deny_unknown_fields)]
+pub struct ListApiKeys {
+    #[deserr(default, error = DeserrQueryParamError<InvalidApiKeyOffset>)]
+    pub offset: Param<usize>,
+    #[deserr(default = Param(PAGINATION_DEFAULT_LIMIT), error = DeserrQueryParamError<InvalidApiKeyLimit>)]
+    pub limit: Param<usize>,
+}
+
+impl ListApiKeys {
+    fn as_pagination(self) -> Pagination {
+        Pagination { offset: self.offset.0, limit: self.limit.0 }
+    }
+}
+
+pub async fn list_api_keys(
+    auth_controller: GuardedData<ActionPolicy<{ actions::KEYS_GET }>, Data<AuthController>>,
+    list_api_keys: AwebQueryParameter<ListApiKeys, DeserrQueryParamError>,
+) -> Result<HttpResponse, ResponseError> {
+    let paginate = list_api_keys.into_inner().as_pagination();
+    let page_view = tokio::task::spawn_blocking(move || -> Result<_, AuthControllerError> {
+        let keys = auth_controller.list_keys()?;
+        let page_view = paginate
+            .auto_paginate_sized(keys.into_iter().map(|k| KeyView::from_key(k, &auth_controller)));
+
+        Ok(page_view)
+    })
+    .await
+    .map_err(|e| ResponseError::from_msg(e.to_string(), Code::Internal))??;
+
+    Ok(HttpResponse::Ok().json(page_view))
+}
+
+pub async fn get_api_key(
+    auth_controller: GuardedData<ActionPolicy<{ actions::KEYS_GET }>, Data<AuthController>>,
+    path: web::Path<AuthParam>,
+) -> Result<HttpResponse, ResponseError> {
+    let key = path.into_inner().key;
+
+    let res = tokio::task::spawn_blocking(move || -> Result<_, AuthControllerError> {
+        let uid =
+            Uuid::parse_str(&key).or_else(|_| auth_controller.get_uid_from_encoded_key(&key))?;
+        let key = auth_controller.get_key(uid)?;
+
+        Ok(KeyView::from_key(key, &auth_controller))
+    })
+    .await
+    .map_err(|e| ResponseError::from_msg(e.to_string(), Code::Internal))??;
+
+    Ok(HttpResponse::Ok().json(res))
+}
+
+pub async fn patch_api_key(
+    auth_controller: GuardedData<ActionPolicy<{ actions::KEYS_UPDATE }>, Data<AuthController>>,
+    body: AwebJson<PatchApiKey, DeserrJsonError>,
+    path: web::Path<AuthParam>,
+) -> Result<HttpResponse, ResponseError> {
+    let key = path.into_inner().key;
+    let patch_api_key = body.into_inner();
+    let res = tokio::task::spawn_blocking(move || -> Result<_, AuthControllerError> {
+        let uid =
+            Uuid::parse_str(&key).or_else(|_| auth_controller.get_uid_from_encoded_key(&key))?;
+        let key = auth_controller.update_key(uid, patch_api_key)?;
+
+        Ok(KeyView::from_key(key, &auth_controller))
+    })
+    .await
+    .map_err(|e| ResponseError::from_msg(e.to_string(), Code::Internal))??;
+
+    Ok(HttpResponse::Ok().json(res))
+}
+
+pub async fn delete_api_key(
+    auth_controller: GuardedData<ActionPolicy<{ actions::KEYS_DELETE }>, Data<AuthController>>,
+    path: web::Path<AuthParam>,
+) -> Result<HttpResponse, ResponseError> {
+    let key = path.into_inner().key;
+    tokio::task::spawn_blocking(move || {
+        let uid =
+            Uuid::parse_str(&key).or_else(|_| auth_controller.get_uid_from_encoded_key(&key))?;
+        auth_controller.delete_key(uid)
+    })
+    .await
+    .map_err(|e| ResponseError::from_msg(e.to_string(), Code::Internal))??;
+
+    Ok(HttpResponse::NoContent().finish())
+}
+
+#[derive(Deserialize)]
+pub struct AuthParam {
+    key: String,
+}
+
+#[derive(Debug, Serialize)]
+#[serde(rename_all = "camelCase")]
+struct KeyView {
+    name: Option<String>,
+    description: Option<String>,
+    key: String,
+    uid: Uuid,
+    actions: Vec<Action>,
+    indexes: Vec<String>,
+    #[serde(serialize_with = "time::serde::rfc3339::option::serialize")]
+    expires_at: Option<OffsetDateTime>,
+    #[serde(serialize_with = "time::serde::rfc3339::serialize")]
+    created_at: OffsetDateTime,
+    #[serde(serialize_with = "time::serde::rfc3339::serialize")]
+    updated_at: OffsetDateTime,
+}
+
+impl KeyView {
+    fn from_key(key: Key, auth: &AuthController) -> Self {
+        let generated_key = auth.generate_key(key.uid).unwrap_or_default();
+
+        KeyView {
+            name: key.name,
+            description: key.description,
+            key: generated_key,
+            uid: key.uid,
+            actions: key.actions,
+            indexes: key.indexes.into_iter().map(|x| x.to_string()).collect(),
+            expires_at: key.expires_at,
+            created_at: key.created_at,
+            updated_at: key.updated_at,
+        }
+    }
+}

crates/meilisearch/src/routes/dump.rs

@@ -0,0 +1,44 @@
+use actix_web::web::Data;
+use actix_web::{web, HttpRequest, HttpResponse};
+use index_scheduler::IndexScheduler;
+use meilisearch_auth::AuthController;
+use meilisearch_types::error::ResponseError;
+use meilisearch_types::tasks::KindWithContent;
+use tracing::debug;
+
+use crate::analytics::Analytics;
+use crate::extractors::authentication::policies::*;
+use crate::extractors::authentication::GuardedData;
+use crate::extractors::sequential_extractor::SeqHandler;
+use crate::routes::{get_task_id, is_dry_run, SummarizedTaskView};
+use crate::Opt;
+
+pub fn configure(cfg: &mut web::ServiceConfig) {
+    cfg.service(web::resource("").route(web::post().to(SeqHandler(create_dump))));
+}
+
+crate::empty_analytics!(DumpAnalytics, "Dump Created");
+
+pub async fn create_dump(
+    index_scheduler: GuardedData<ActionPolicy<{ actions::DUMPS_CREATE }>, Data<IndexScheduler>>,
+    auth_controller: GuardedData<ActionPolicy<{ actions::DUMPS_CREATE }>, Data<AuthController>>,
+    req: HttpRequest,
+    opt: web::Data<Opt>,
+    analytics: web::Data<Analytics>,
+) -> Result<HttpResponse, ResponseError> {
+    analytics.publish(DumpAnalytics::default(), &req);
+
+    let task = KindWithContent::DumpCreation {
+        keys: auth_controller.list_keys()?,
+        instance_uid: analytics.instance_uid().cloned(),
+    };
+    let uid = get_task_id(&req, &opt)?;
+    let dry_run = is_dry_run(&req, &opt)?;
+    let task: SummarizedTaskView =
+        tokio::task::spawn_blocking(move || index_scheduler.register(task, uid, dry_run))
+            .await??
+            .into();
+
+    debug!(returns = ?task, "Create dump");
+    Ok(HttpResponse::Accepted().json(task))
+}

crates/meilisearch/src/routes/features.rs

@@ -0,0 +1,130 @@
+use actix_web::web::{self, Data};
+use actix_web::{HttpRequest, HttpResponse};
+use deserr::actix_web::AwebJson;
+use deserr::Deserr;
+use index_scheduler::IndexScheduler;
+use meilisearch_types::deserr::DeserrJsonError;
+use meilisearch_types::error::ResponseError;
+use meilisearch_types::keys::actions;
+use serde::Serialize;
+use tracing::debug;
+
+use crate::analytics::{Aggregate, Analytics};
+use crate::extractors::authentication::policies::ActionPolicy;
+use crate::extractors::authentication::GuardedData;
+use crate::extractors::sequential_extractor::SeqHandler;
+
+pub fn configure(cfg: &mut web::ServiceConfig) {
+    cfg.service(
+        web::resource("")
+            .route(web::get().to(get_features))
+            .route(web::patch().to(SeqHandler(patch_features))),
+    );
+}
+
+async fn get_features(
+    index_scheduler: GuardedData<
+        ActionPolicy<{ actions::EXPERIMENTAL_FEATURES_GET }>,
+        Data<IndexScheduler>,
+    >,
+) -> HttpResponse {
+    let features = index_scheduler.features();
+
+    let features = features.runtime_features();
+    debug!(returns = ?features, "Get features");
+    HttpResponse::Ok().json(features)
+}
+
+#[derive(Debug, Deserr)]
+#[deserr(error = DeserrJsonError, rename_all = camelCase, deny_unknown_fields)]
+pub struct RuntimeTogglableFeatures {
+    #[deserr(default)]
+    pub vector_store: Option<bool>,
+    #[deserr(default)]
+    pub metrics: Option<bool>,
+    #[deserr(default)]
+    pub logs_route: Option<bool>,
+    #[deserr(default)]
+    pub edit_documents_by_function: Option<bool>,
+    #[deserr(default)]
+    pub contains_filter: Option<bool>,
+}
+
+#[derive(Serialize)]
+pub struct PatchExperimentalFeatureAnalytics {
+    vector_store: bool,
+    metrics: bool,
+    logs_route: bool,
+    edit_documents_by_function: bool,
+    contains_filter: bool,
+}
+
+impl Aggregate for PatchExperimentalFeatureAnalytics {
+    fn event_name(&self) -> &'static str {
+        "Experimental features Updated"
+    }
+
+    fn aggregate(self: Box<Self>, new: Box<Self>) -> Box<Self> {
+        Box::new(Self {
+            vector_store: new.vector_store,
+            metrics: new.metrics,
+            logs_route: new.logs_route,
+            edit_documents_by_function: new.edit_documents_by_function,
+            contains_filter: new.contains_filter,
+        })
+    }
+
+    fn into_event(self: Box<Self>) -> serde_json::Value {
+        serde_json::to_value(*self).unwrap_or_default()
+    }
+}
+
+async fn patch_features(
+    index_scheduler: GuardedData<
+        ActionPolicy<{ actions::EXPERIMENTAL_FEATURES_UPDATE }>,
+        Data<IndexScheduler>,
+    >,
+    new_features: AwebJson<RuntimeTogglableFeatures, DeserrJsonError>,
+    req: HttpRequest,
+    analytics: Data<Analytics>,
+) -> Result<HttpResponse, ResponseError> {
+    let features = index_scheduler.features();
+    debug!(parameters = ?new_features, "Patch features");
+
+    let old_features = features.runtime_features();
+    let new_features = meilisearch_types::features::RuntimeTogglableFeatures {
+        vector_store: new_features.0.vector_store.unwrap_or(old_features.vector_store),
+        metrics: new_features.0.metrics.unwrap_or(old_features.metrics),
+        logs_route: new_features.0.logs_route.unwrap_or(old_features.logs_route),
+        edit_documents_by_function: new_features
+            .0
+            .edit_documents_by_function
+            .unwrap_or(old_features.edit_documents_by_function),
+        contains_filter: new_features.0.contains_filter.unwrap_or(old_features.contains_filter),
+    };
+
+    // explicitly destructure for analytics rather than using the `Serialize` implementation, because
+    // the it renames to camelCase, which we don't want for analytics.
+    // **Do not** ignore fields with `..` or `_` here, because we want to add them in the future.
+    let meilisearch_types::features::RuntimeTogglableFeatures {
+        vector_store,
+        metrics,
+        logs_route,
+        edit_documents_by_function,
+        contains_filter,
+    } = new_features;
+
+    analytics.publish(
+        PatchExperimentalFeatureAnalytics {
+            vector_store,
+            metrics,
+            logs_route,
+            edit_documents_by_function,
+            contains_filter,
+        },
+        &req,
+    );
+    index_scheduler.put_runtime_features(new_features)?;
+    debug!(returns = ?new_features, "Patch features");
+    Ok(HttpResponse::Ok().json(new_features))
+}

crates/meilisearch/src/routes/indexes/facet_search.rs

@@ -0,0 +1,260 @@
+use std::collections::{BinaryHeap, HashSet};
+
+use actix_web::web::Data;
+use actix_web::{web, HttpRequest, HttpResponse};
+use deserr::actix_web::AwebJson;
+use index_scheduler::IndexScheduler;
+use meilisearch_types::deserr::DeserrJsonError;
+use meilisearch_types::error::deserr_codes::*;
+use meilisearch_types::error::ResponseError;
+use meilisearch_types::index_uid::IndexUid;
+use meilisearch_types::locales::Locale;
+use serde_json::Value;
+use tracing::debug;
+
+use crate::analytics::{Aggregate, Analytics};
+use crate::extractors::authentication::policies::*;
+use crate::extractors::authentication::GuardedData;
+use crate::routes::indexes::search::search_kind;
+use crate::search::{
+    add_search_rules, perform_facet_search, FacetSearchResult, HybridQuery, MatchingStrategy,
+    RankingScoreThreshold, SearchQuery, DEFAULT_CROP_LENGTH, DEFAULT_CROP_MARKER,
+    DEFAULT_HIGHLIGHT_POST_TAG, DEFAULT_HIGHLIGHT_PRE_TAG, DEFAULT_SEARCH_LIMIT,
+    DEFAULT_SEARCH_OFFSET,
+};
+use crate::search_queue::SearchQueue;
+
+pub fn configure(cfg: &mut web::ServiceConfig) {
+    cfg.service(web::resource("").route(web::post().to(search)));
+}
+
+/// # Important
+///
+/// Intentionally don't use `deny_unknown_fields` to ignore search parameters sent by user
+#[derive(Debug, Clone, Default, PartialEq, deserr::Deserr)]
+#[deserr(error = DeserrJsonError, rename_all = camelCase)]
+pub struct FacetSearchQuery {
+    #[deserr(default, error = DeserrJsonError<InvalidFacetSearchQuery>)]
+    pub facet_query: Option<String>,
+    #[deserr(error = DeserrJsonError<InvalidFacetSearchFacetName>, missing_field_error = DeserrJsonError::missing_facet_search_facet_name)]
+    pub facet_name: String,
+    #[deserr(default, error = DeserrJsonError<InvalidSearchQ>)]
+    pub q: Option<String>,
+    #[deserr(default, error = DeserrJsonError<InvalidSearchVector>)]
+    pub vector: Option<Vec<f32>>,
+    #[deserr(default, error = DeserrJsonError<InvalidHybridQuery>)]
+    pub hybrid: Option<HybridQuery>,
+    #[deserr(default, error = DeserrJsonError<InvalidSearchFilter>)]
+    pub filter: Option<Value>,
+    #[deserr(default, error = DeserrJsonError<InvalidSearchMatchingStrategy>, default)]
+    pub matching_strategy: MatchingStrategy,
+    #[deserr(default, error = DeserrJsonError<InvalidSearchAttributesToSearchOn>, default)]
+    pub attributes_to_search_on: Option<Vec<String>>,
+    #[deserr(default, error = DeserrJsonError<InvalidSearchRankingScoreThreshold>, default)]
+    pub ranking_score_threshold: Option<RankingScoreThreshold>,
+    #[deserr(default, error = DeserrJsonError<InvalidSearchLocales>, default)]
+    pub locales: Option<Vec<Locale>>,
+}
+
+#[derive(Default)]
+pub struct FacetSearchAggregator {
+    // requests
+    total_received: usize,
+    total_succeeded: usize,
+    time_spent: BinaryHeap<usize>,
+
+    // The set of all facetNames that were used
+    facet_names: HashSet<String>,
+
+    // As there been any other parameter than the facetName or facetQuery ones?
+    additional_search_parameters_provided: bool,
+}
+
+impl FacetSearchAggregator {
+    #[allow(clippy::field_reassign_with_default)]
+    pub fn from_query(query: &FacetSearchQuery) -> Self {
+        let FacetSearchQuery {
+            facet_query: _,
+            facet_name,
+            vector,
+            q,
+            filter,
+            matching_strategy,
+            attributes_to_search_on,
+            hybrid,
+            ranking_score_threshold,
+            locales,
+        } = query;
+
+        Self {
+            total_received: 1,
+            facet_names: Some(facet_name.clone()).into_iter().collect(),
+            additional_search_parameters_provided: q.is_some()
+                || vector.is_some()
+                || filter.is_some()
+                || *matching_strategy != MatchingStrategy::default()
+                || attributes_to_search_on.is_some()
+                || hybrid.is_some()
+                || ranking_score_threshold.is_some()
+                || locales.is_some(),
+            ..Default::default()
+        }
+    }
+
+    pub fn succeed(&mut self, result: &FacetSearchResult) {
+        let FacetSearchResult { facet_hits: _, facet_query: _, processing_time_ms } = result;
+        self.total_succeeded = 1;
+        self.time_spent.push(*processing_time_ms as usize);
+    }
+}
+
+impl Aggregate for FacetSearchAggregator {
+    fn event_name(&self) -> &'static str {
+        "Facet Searched POST"
+    }
+
+    fn aggregate(mut self: Box<Self>, new: Box<Self>) -> Box<Self> {
+        for time in new.time_spent {
+            self.time_spent.push(time);
+        }
+
+        Box::new(Self {
+            total_received: self.total_received.saturating_add(new.total_received),
+            total_succeeded: self.total_succeeded.saturating_add(new.total_succeeded),
+            time_spent: self.time_spent,
+            facet_names: self.facet_names.union(&new.facet_names).cloned().collect(),
+            additional_search_parameters_provided: self.additional_search_parameters_provided
+                | new.additional_search_parameters_provided,
+        })
+    }
+
+    fn into_event(self: Box<Self>) -> serde_json::Value {
+        let Self {
+            total_received,
+            total_succeeded,
+            time_spent,
+            facet_names,
+            additional_search_parameters_provided,
+        } = *self;
+        // the index of the 99th percentage of value
+        let percentile_99th = 0.99 * (total_succeeded as f64 - 1.) + 1.;
+        // we get all the values in a sorted manner
+        let time_spent = time_spent.into_sorted_vec();
+        // We are only interested by the slowest value of the 99th fastest results
+        let time_spent = time_spent.get(percentile_99th as usize);
+
+        serde_json::json!({
+            "requests": {
+                "99th_response_time":  time_spent.map(|t| format!("{:.2}", t)),
+                "total_succeeded": total_succeeded,
+                "total_failed": total_received.saturating_sub(total_succeeded), // just to be sure we never panics
+                "total_received": total_received,
+            },
+            "facets": {
+                "total_distinct_facet_count": facet_names.len(),
+                "additional_search_parameters_provided": additional_search_parameters_provided,
+            },
+        })
+    }
+}
+
+pub async fn search(
+    index_scheduler: GuardedData<ActionPolicy<{ actions::SEARCH }>, Data<IndexScheduler>>,
+    search_queue: Data<SearchQueue>,
+    index_uid: web::Path<String>,
+    params: AwebJson<FacetSearchQuery, DeserrJsonError>,
+    req: HttpRequest,
+    analytics: web::Data<Analytics>,
+) -> Result<HttpResponse, ResponseError> {
+    let index_uid = IndexUid::try_from(index_uid.into_inner())?;
+
+    let query = params.into_inner();
+    debug!(parameters = ?query, "Facet search");
+
+    let mut aggregate = FacetSearchAggregator::from_query(&query);
+
+    let facet_query = query.facet_query.clone();
+    let facet_name = query.facet_name.clone();
+    let locales = query.locales.clone().map(|l| l.into_iter().map(Into::into).collect());
+    let mut search_query = SearchQuery::from(query);
+
+    // Tenant token search_rules.
+    if let Some(search_rules) = index_scheduler.filters().get_index_search_rules(&index_uid) {
+        add_search_rules(&mut search_query.filter, search_rules);
+    }
+
+    let index = index_scheduler.index(&index_uid)?;
+    let features = index_scheduler.features();
+    let search_kind = search_kind(&search_query, &index_scheduler, &index, features)?;
+    let permit = search_queue.try_get_search_permit().await?;
+    let search_result = tokio::task::spawn_blocking(move || {
+        perform_facet_search(
+            &index,
+            search_query,
+            facet_query,
+            facet_name,
+            search_kind,
+            index_scheduler.features(),
+            locales,
+        )
+    })
+    .await;
+    permit.drop().await;
+    let search_result = search_result?;
+
+    if let Ok(ref search_result) = search_result {
+        aggregate.succeed(search_result);
+    }
+    analytics.publish(aggregate, &req);
+
+    let search_result = search_result?;
+
+    debug!(returns = ?search_result, "Facet search");
+    Ok(HttpResponse::Ok().json(search_result))
+}
+
+impl From<FacetSearchQuery> for SearchQuery {
+    fn from(value: FacetSearchQuery) -> Self {
+        let FacetSearchQuery {
+            facet_query: _,
+            facet_name: _,
+            q,
+            vector,
+            filter,
+            matching_strategy,
+            attributes_to_search_on,
+            hybrid,
+            ranking_score_threshold,
+            locales,
+        } = value;
+
+        SearchQuery {
+            q,
+            offset: DEFAULT_SEARCH_OFFSET(),
+            limit: DEFAULT_SEARCH_LIMIT(),
+            page: None,
+            hits_per_page: None,
+            attributes_to_retrieve: None,
+            retrieve_vectors: false,
+            attributes_to_crop: None,
+            crop_length: DEFAULT_CROP_LENGTH(),
+            attributes_to_highlight: None,
+            show_matches_position: false,
+            show_ranking_score: false,
+            show_ranking_score_details: false,
+            filter,
+            sort: None,
+            distinct: None,
+            facets: None,
+            highlight_pre_tag: DEFAULT_HIGHLIGHT_PRE_TAG(),
+            highlight_post_tag: DEFAULT_HIGHLIGHT_POST_TAG(),
+            crop_marker: DEFAULT_CROP_MARKER(),
+            matching_strategy,
+            vector,
+            attributes_to_search_on,
+            hybrid,
+            ranking_score_threshold,
+            locales,
+        }
+    }
+}

crates/meilisearch/src/routes/indexes/mod.rs

@@ -0,0 +1,318 @@
+use std::collections::BTreeSet;
+use std::convert::Infallible;
+
+use actix_web::web::Data;
+use actix_web::{web, HttpRequest, HttpResponse};
+use deserr::actix_web::{AwebJson, AwebQueryParameter};
+use deserr::{DeserializeError, Deserr, ValuePointerRef};
+use index_scheduler::IndexScheduler;
+use meilisearch_types::deserr::query_params::Param;
+use meilisearch_types::deserr::{immutable_field_error, DeserrJsonError, DeserrQueryParamError};
+use meilisearch_types::error::deserr_codes::*;
+use meilisearch_types::error::{Code, ResponseError};
+use meilisearch_types::index_uid::IndexUid;
+use meilisearch_types::milli::{self, FieldDistribution, Index};
+use meilisearch_types::tasks::KindWithContent;
+use serde::Serialize;
+use time::OffsetDateTime;
+use tracing::debug;
+
+use super::{get_task_id, Pagination, SummarizedTaskView, PAGINATION_DEFAULT_LIMIT};
+use crate::analytics::{Aggregate, Analytics};
+use crate::extractors::authentication::policies::*;
+use crate::extractors::authentication::{AuthenticationError, GuardedData};
+use crate::extractors::sequential_extractor::SeqHandler;
+use crate::routes::is_dry_run;
+use crate::Opt;
+
+pub mod documents;
+pub mod facet_search;
+pub mod search;
+mod search_analytics;
+pub mod settings;
+mod settings_analytics;
+pub mod similar;
+mod similar_analytics;
+
+pub fn configure(cfg: &mut web::ServiceConfig) {
+    cfg.service(
+        web::resource("")
+            .route(web::get().to(list_indexes))
+            .route(web::post().to(SeqHandler(create_index))),
+    )
+    .service(
+        web::scope("/{index_uid}")
+            .service(
+                web::resource("")
+                    .route(web::get().to(SeqHandler(get_index)))
+                    .route(web::patch().to(SeqHandler(update_index)))
+                    .route(web::delete().to(SeqHandler(delete_index))),
+            )
+            .service(web::resource("/stats").route(web::get().to(SeqHandler(get_index_stats))))
+            .service(web::scope("/documents").configure(documents::configure))
+            .service(web::scope("/search").configure(search::configure))
+            .service(web::scope("/facet-search").configure(facet_search::configure))
+            .service(web::scope("/similar").configure(similar::configure))
+            .service(web::scope("/settings").configure(settings::configure)),
+    );
+}
+
+#[derive(Debug, Serialize, Clone)]
+#[serde(rename_all = "camelCase")]
+pub struct IndexView {
+    pub uid: String,
+    #[serde(with = "time::serde::rfc3339")]
+    pub created_at: OffsetDateTime,
+    #[serde(with = "time::serde::rfc3339")]
+    pub updated_at: OffsetDateTime,
+    pub primary_key: Option<String>,
+}
+
+impl IndexView {
+    fn new(uid: String, index: &Index) -> Result<IndexView, milli::Error> {
+        // It is important that this function does not keep the Index handle or a clone of it, because
+        // `list_indexes` relies on this property to avoid opening all indexes at once.
+        let rtxn = index.read_txn()?;
+        Ok(IndexView {
+            uid,
+            created_at: index.created_at(&rtxn)?,
+            updated_at: index.updated_at(&rtxn)?,
+            primary_key: index.primary_key(&rtxn)?.map(String::from),
+        })
+    }
+}
+
+#[derive(Deserr, Debug, Clone, Copy)]
+#[deserr(error = DeserrQueryParamError, rename_all = camelCase, deny_unknown_fields)]
+pub struct ListIndexes {
+    #[deserr(default, error = DeserrQueryParamError<InvalidIndexOffset>)]
+    pub offset: Param<usize>,
+    #[deserr(default = Param(PAGINATION_DEFAULT_LIMIT), error = DeserrQueryParamError<InvalidIndexLimit>)]
+    pub limit: Param<usize>,
+}
+impl ListIndexes {
+    fn as_pagination(self) -> Pagination {
+        Pagination { offset: self.offset.0, limit: self.limit.0 }
+    }
+}
+
+pub async fn list_indexes(
+    index_scheduler: GuardedData<ActionPolicy<{ actions::INDEXES_GET }>, Data<IndexScheduler>>,
+    paginate: AwebQueryParameter<ListIndexes, DeserrQueryParamError>,
+) -> Result<HttpResponse, ResponseError> {
+    debug!(parameters = ?paginate, "List indexes");
+    let filters = index_scheduler.filters();
+    let indexes: Vec<Option<IndexView>> =
+        index_scheduler.try_for_each_index(|uid, index| -> Result<Option<IndexView>, _> {
+            if !filters.is_index_authorized(uid) {
+                return Ok(None);
+            }
+            Ok(Some(IndexView::new(uid.to_string(), index)?))
+        })?;
+    // Won't cause to open all indexes because IndexView doesn't keep the `Index` opened.
+    let indexes: Vec<IndexView> = indexes.into_iter().flatten().collect();
+    let ret = paginate.as_pagination().auto_paginate_sized(indexes.into_iter());
+
+    debug!(returns = ?ret, "List indexes");
+    Ok(HttpResponse::Ok().json(ret))
+}
+
+#[derive(Deserr, Debug)]
+#[deserr(error = DeserrJsonError, rename_all = camelCase, deny_unknown_fields)]
+pub struct IndexCreateRequest {
+    #[deserr(error = DeserrJsonError<InvalidIndexUid>, missing_field_error = DeserrJsonError::missing_index_uid)]
+    uid: IndexUid,
+    #[deserr(default, error = DeserrJsonError<InvalidIndexPrimaryKey>)]
+    primary_key: Option<String>,
+}
+
+#[derive(Serialize)]
+struct IndexCreatedAggregate {
+    primary_key: BTreeSet<String>,
+}
+
+impl Aggregate for IndexCreatedAggregate {
+    fn event_name(&self) -> &'static str {
+        "Index Created"
+    }
+
+    fn aggregate(self: Box<Self>, new: Box<Self>) -> Box<Self> {
+        Box::new(Self { primary_key: self.primary_key.union(&new.primary_key).cloned().collect() })
+    }
+
+    fn into_event(self: Box<Self>) -> serde_json::Value {
+        serde_json::to_value(*self).unwrap_or_default()
+    }
+}
+
+pub async fn create_index(
+    index_scheduler: GuardedData<ActionPolicy<{ actions::INDEXES_CREATE }>, Data<IndexScheduler>>,
+    body: AwebJson<IndexCreateRequest, DeserrJsonError>,
+    req: HttpRequest,
+    opt: web::Data<Opt>,
+    analytics: web::Data<Analytics>,
+) -> Result<HttpResponse, ResponseError> {
+    debug!(parameters = ?body, "Create index");
+    let IndexCreateRequest { primary_key, uid } = body.into_inner();
+
+    let allow_index_creation = index_scheduler.filters().allow_index_creation(&uid);
+    if allow_index_creation {
+        analytics.publish(
+            IndexCreatedAggregate { primary_key: primary_key.iter().cloned().collect() },
+            &req,
+        );
+
+        let task = KindWithContent::IndexCreation { index_uid: uid.to_string(), primary_key };
+        let uid = get_task_id(&req, &opt)?;
+        let dry_run = is_dry_run(&req, &opt)?;
+        let task: SummarizedTaskView =
+            tokio::task::spawn_blocking(move || index_scheduler.register(task, uid, dry_run))
+                .await??
+                .into();
+        debug!(returns = ?task, "Create index");
+
+        Ok(HttpResponse::Accepted().json(task))
+    } else {
+        Err(AuthenticationError::InvalidToken.into())
+    }
+}
+
+fn deny_immutable_fields_index(
+    field: &str,
+    accepted: &[&str],
+    location: ValuePointerRef,
+) -> DeserrJsonError {
+    match field {
+        "uid" => immutable_field_error(field, accepted, Code::ImmutableIndexUid),
+        "createdAt" => immutable_field_error(field, accepted, Code::ImmutableIndexCreatedAt),
+        "updatedAt" => immutable_field_error(field, accepted, Code::ImmutableIndexUpdatedAt),
+        _ => deserr::take_cf_content(DeserrJsonError::<BadRequest>::error::<Infallible>(
+            None,
+            deserr::ErrorKind::UnknownKey { key: field, accepted },
+            location,
+        )),
+    }
+}
+
+#[derive(Deserr, Debug)]
+#[deserr(error = DeserrJsonError, rename_all = camelCase, deny_unknown_fields = deny_immutable_fields_index)]
+pub struct UpdateIndexRequest {
+    #[deserr(default, error = DeserrJsonError<InvalidIndexPrimaryKey>)]
+    primary_key: Option<String>,
+}
+
+pub async fn get_index(
+    index_scheduler: GuardedData<ActionPolicy<{ actions::INDEXES_GET }>, Data<IndexScheduler>>,
+    index_uid: web::Path<String>,
+) -> Result<HttpResponse, ResponseError> {
+    let index_uid = IndexUid::try_from(index_uid.into_inner())?;
+
+    let index = index_scheduler.index(&index_uid)?;
+    let index_view = IndexView::new(index_uid.into_inner(), &index)?;
+
+    debug!(returns = ?index_view, "Get index");
+
+    Ok(HttpResponse::Ok().json(index_view))
+}
+
+#[derive(Serialize)]
+struct IndexUpdatedAggregate {
+    primary_key: BTreeSet<String>,
+}
+
+impl Aggregate for IndexUpdatedAggregate {
+    fn event_name(&self) -> &'static str {
+        "Index Updated"
+    }
+
+    fn aggregate(self: Box<Self>, new: Box<Self>) -> Box<Self> {
+        Box::new(Self { primary_key: self.primary_key.union(&new.primary_key).cloned().collect() })
+    }
+
+    fn into_event(self: Box<Self>) -> serde_json::Value {
+        serde_json::to_value(*self).unwrap_or_default()
+    }
+}
+pub async fn update_index(
+    index_scheduler: GuardedData<ActionPolicy<{ actions::INDEXES_UPDATE }>, Data<IndexScheduler>>,
+    index_uid: web::Path<String>,
+    body: AwebJson<UpdateIndexRequest, DeserrJsonError>,
+    req: HttpRequest,
+    opt: web::Data<Opt>,
+    analytics: web::Data<Analytics>,
+) -> Result<HttpResponse, ResponseError> {
+    debug!(parameters = ?body, "Update index");
+    let index_uid = IndexUid::try_from(index_uid.into_inner())?;
+    let body = body.into_inner();
+    analytics.publish(
+        IndexUpdatedAggregate { primary_key: body.primary_key.iter().cloned().collect() },
+        &req,
+    );
+
+    let task = KindWithContent::IndexUpdate {
+        index_uid: index_uid.into_inner(),
+        primary_key: body.primary_key,
+    };
+
+    let uid = get_task_id(&req, &opt)?;
+    let dry_run = is_dry_run(&req, &opt)?;
+    let task: SummarizedTaskView =
+        tokio::task::spawn_blocking(move || index_scheduler.register(task, uid, dry_run))
+            .await??
+            .into();
+
+    debug!(returns = ?task, "Update index");
+    Ok(HttpResponse::Accepted().json(task))
+}
+
+pub async fn delete_index(
+    index_scheduler: GuardedData<ActionPolicy<{ actions::INDEXES_DELETE }>, Data<IndexScheduler>>,
+    index_uid: web::Path<String>,
+    req: HttpRequest,
+    opt: web::Data<Opt>,
+) -> Result<HttpResponse, ResponseError> {
+    let index_uid = IndexUid::try_from(index_uid.into_inner())?;
+    let task = KindWithContent::IndexDeletion { index_uid: index_uid.into_inner() };
+    let uid = get_task_id(&req, &opt)?;
+    let dry_run = is_dry_run(&req, &opt)?;
+    let task: SummarizedTaskView =
+        tokio::task::spawn_blocking(move || index_scheduler.register(task, uid, dry_run))
+            .await??
+            .into();
+    debug!(returns = ?task, "Delete index");
+
+    Ok(HttpResponse::Accepted().json(task))
+}
+
+/// Stats of an `Index`, as known to the `stats` route.
+#[derive(Serialize, Debug)]
+#[serde(rename_all = "camelCase")]
+pub struct IndexStats {
+    /// Number of documents in the index
+    pub number_of_documents: u64,
+    /// Whether the index is currently performing indexation, according to the scheduler.
+    pub is_indexing: bool,
+    /// Association of every field name with the number of times it occurs in the documents.
+    pub field_distribution: FieldDistribution,
+}
+
+impl From<index_scheduler::IndexStats> for IndexStats {
+    fn from(stats: index_scheduler::IndexStats) -> Self {
+        IndexStats {
+            number_of_documents: stats.inner_stats.number_of_documents,
+            is_indexing: stats.is_indexing,
+            field_distribution: stats.inner_stats.field_distribution,
+        }
+    }
+}
+
+pub async fn get_index_stats(
+    index_scheduler: GuardedData<ActionPolicy<{ actions::STATS_GET }>, Data<IndexScheduler>>,
+    index_uid: web::Path<String>,
+) -> Result<HttpResponse, ResponseError> {
+    let index_uid = IndexUid::try_from(index_uid.into_inner())?;
+    let stats = IndexStats::from(index_scheduler.index_stats(&index_uid)?);
+
+    debug!(returns = ?stats, "Get index stats");
+    Ok(HttpResponse::Ok().json(stats))
+}

crates/meilisearch/src/routes/indexes/search.rs

@@ -0,0 +1,385 @@
+use actix_web::web::Data;
+use actix_web::{web, HttpRequest, HttpResponse};
+use deserr::actix_web::{AwebJson, AwebQueryParameter};
+use index_scheduler::{IndexScheduler, RoFeatures};
+use meilisearch_types::deserr::query_params::Param;
+use meilisearch_types::deserr::{DeserrJsonError, DeserrQueryParamError};
+use meilisearch_types::error::deserr_codes::*;
+use meilisearch_types::error::ResponseError;
+use meilisearch_types::index_uid::IndexUid;
+use meilisearch_types::locales::Locale;
+use meilisearch_types::milli;
+use meilisearch_types::serde_cs::vec::CS;
+use serde_json::Value;
+use tracing::debug;
+
+use crate::analytics::Analytics;
+use crate::error::MeilisearchHttpError;
+use crate::extractors::authentication::policies::*;
+use crate::extractors::authentication::GuardedData;
+use crate::extractors::sequential_extractor::SeqHandler;
+use crate::metrics::MEILISEARCH_DEGRADED_SEARCH_REQUESTS;
+use crate::routes::indexes::search_analytics::{SearchAggregator, SearchGET, SearchPOST};
+use crate::search::{
+    add_search_rules, perform_search, HybridQuery, MatchingStrategy, RankingScoreThreshold,
+    RetrieveVectors, SearchKind, SearchQuery, SemanticRatio, DEFAULT_CROP_LENGTH,
+    DEFAULT_CROP_MARKER, DEFAULT_HIGHLIGHT_POST_TAG, DEFAULT_HIGHLIGHT_PRE_TAG,
+    DEFAULT_SEARCH_LIMIT, DEFAULT_SEARCH_OFFSET, DEFAULT_SEMANTIC_RATIO,
+};
+use crate::search_queue::SearchQueue;
+
+pub fn configure(cfg: &mut web::ServiceConfig) {
+    cfg.service(
+        web::resource("")
+            .route(web::get().to(SeqHandler(search_with_url_query)))
+            .route(web::post().to(SeqHandler(search_with_post))),
+    );
+}
+
+#[derive(Debug, deserr::Deserr)]
+#[deserr(error = DeserrQueryParamError, rename_all = camelCase, deny_unknown_fields)]
+pub struct SearchQueryGet {
+    #[deserr(default, error = DeserrQueryParamError<InvalidSearchQ>)]
+    q: Option<String>,
+    #[deserr(default, error = DeserrQueryParamError<InvalidSearchVector>)]
+    vector: Option<CS<f32>>,
+    #[deserr(default = Param(DEFAULT_SEARCH_OFFSET()), error = DeserrQueryParamError<InvalidSearchOffset>)]
+    offset: Param<usize>,
+    #[deserr(default = Param(DEFAULT_SEARCH_LIMIT()), error = DeserrQueryParamError<InvalidSearchLimit>)]
+    limit: Param<usize>,
+    #[deserr(default, error = DeserrQueryParamError<InvalidSearchPage>)]
+    page: Option<Param<usize>>,
+    #[deserr(default, error = DeserrQueryParamError<InvalidSearchHitsPerPage>)]
+    hits_per_page: Option<Param<usize>>,
+    #[deserr(default, error = DeserrQueryParamError<InvalidSearchAttributesToRetrieve>)]
+    attributes_to_retrieve: Option<CS<String>>,
+    #[deserr(default, error = DeserrQueryParamError<InvalidSearchRetrieveVectors>)]
+    retrieve_vectors: Param<bool>,
+    #[deserr(default, error = DeserrQueryParamError<InvalidSearchAttributesToCrop>)]
+    attributes_to_crop: Option<CS<String>>,
+    #[deserr(default = Param(DEFAULT_CROP_LENGTH()), error = DeserrQueryParamError<InvalidSearchCropLength>)]
+    crop_length: Param<usize>,
+    #[deserr(default, error = DeserrQueryParamError<InvalidSearchAttributesToHighlight>)]
+    attributes_to_highlight: Option<CS<String>>,
+    #[deserr(default, error = DeserrQueryParamError<InvalidSearchFilter>)]
+    filter: Option<String>,
+    #[deserr(default, error = DeserrQueryParamError<InvalidSearchSort>)]
+    sort: Option<String>,
+    #[deserr(default, error = DeserrQueryParamError<InvalidSearchDistinct>)]
+    distinct: Option<String>,
+    #[deserr(default, error = DeserrQueryParamError<InvalidSearchShowMatchesPosition>)]
+    show_matches_position: Param<bool>,
+    #[deserr(default, error = DeserrQueryParamError<InvalidSearchShowRankingScore>)]
+    show_ranking_score: Param<bool>,
+    #[deserr(default, error = DeserrQueryParamError<InvalidSearchShowRankingScoreDetails>)]
+    show_ranking_score_details: Param<bool>,
+    #[deserr(default, error = DeserrQueryParamError<InvalidSearchFacets>)]
+    facets: Option<CS<String>>,
+    #[deserr( default = DEFAULT_HIGHLIGHT_PRE_TAG(), error = DeserrQueryParamError<InvalidSearchHighlightPreTag>)]
+    highlight_pre_tag: String,
+    #[deserr( default = DEFAULT_HIGHLIGHT_POST_TAG(), error = DeserrQueryParamError<InvalidSearchHighlightPostTag>)]
+    highlight_post_tag: String,
+    #[deserr(default = DEFAULT_CROP_MARKER(), error = DeserrQueryParamError<InvalidSearchCropMarker>)]
+    crop_marker: String,
+    #[deserr(default, error = DeserrQueryParamError<InvalidSearchMatchingStrategy>)]
+    matching_strategy: MatchingStrategy,
+    #[deserr(default, error = DeserrQueryParamError<InvalidSearchAttributesToSearchOn>)]
+    pub attributes_to_search_on: Option<CS<String>>,
+    #[deserr(default, error = DeserrQueryParamError<InvalidEmbedder>)]
+    pub hybrid_embedder: Option<String>,
+    #[deserr(default, error = DeserrQueryParamError<InvalidSearchSemanticRatio>)]
+    pub hybrid_semantic_ratio: Option<SemanticRatioGet>,
+    #[deserr(default, error = DeserrQueryParamError<InvalidSearchRankingScoreThreshold>)]
+    pub ranking_score_threshold: Option<RankingScoreThresholdGet>,
+    #[deserr(default, error = DeserrQueryParamError<InvalidSearchLocales>)]
+    pub locales: Option<CS<Locale>>,
+}
+
+#[derive(Debug, Clone, Copy, PartialEq, deserr::Deserr)]
+#[deserr(try_from(String) = TryFrom::try_from -> InvalidSearchRankingScoreThreshold)]
+pub struct RankingScoreThresholdGet(RankingScoreThreshold);
+
+impl std::convert::TryFrom<String> for RankingScoreThresholdGet {
+    type Error = InvalidSearchRankingScoreThreshold;
+
+    fn try_from(s: String) -> Result<Self, Self::Error> {
+        let f: f64 = s.parse().map_err(|_| InvalidSearchRankingScoreThreshold)?;
+        Ok(RankingScoreThresholdGet(RankingScoreThreshold::try_from(f)?))
+    }
+}
+
+#[derive(Debug, Clone, Copy, Default, PartialEq, deserr::Deserr)]
+#[deserr(try_from(String) = TryFrom::try_from -> InvalidSearchSemanticRatio)]
+pub struct SemanticRatioGet(SemanticRatio);
+
+impl std::convert::TryFrom<String> for SemanticRatioGet {
+    type Error = InvalidSearchSemanticRatio;
+
+    fn try_from(s: String) -> Result<Self, Self::Error> {
+        let f: f32 = s.parse().map_err(|_| InvalidSearchSemanticRatio)?;
+        Ok(SemanticRatioGet(SemanticRatio::try_from(f)?))
+    }
+}
+
+impl std::ops::Deref for SemanticRatioGet {
+    type Target = SemanticRatio;
+
+    fn deref(&self) -> &Self::Target {
+        &self.0
+    }
+}
+
+impl TryFrom<SearchQueryGet> for SearchQuery {
+    type Error = ResponseError;
+
+    fn try_from(other: SearchQueryGet) -> Result<Self, Self::Error> {
+        let filter = match other.filter {
+            Some(f) => match serde_json::from_str(&f) {
+                Ok(v) => Some(v),
+                _ => Some(Value::String(f)),
+            },
+            None => None,
+        };
+
+        let hybrid = match (other.hybrid_embedder, other.hybrid_semantic_ratio) {
+            (None, None) => None,
+            (None, Some(_)) => {
+                return Err(ResponseError::from_msg(
+                    "`hybridEmbedder` is mandatory when `hybridSemanticRatio` is present".into(),
+                    meilisearch_types::error::Code::InvalidHybridQuery,
+                ));
+            }
+            (Some(embedder), None) => {
+                Some(HybridQuery { semantic_ratio: DEFAULT_SEMANTIC_RATIO(), embedder })
+            }
+            (Some(embedder), Some(semantic_ratio)) => {
+                Some(HybridQuery { semantic_ratio: *semantic_ratio, embedder })
+            }
+        };
+
+        if other.vector.is_some() && hybrid.is_none() {
+            return Err(ResponseError::from_msg(
+                "`hybridEmbedder` is mandatory when `vector` is present".into(),
+                meilisearch_types::error::Code::MissingSearchHybrid,
+            ));
+        }
+
+        Ok(Self {
+            q: other.q,
+            vector: other.vector.map(CS::into_inner),
+            offset: other.offset.0,
+            limit: other.limit.0,
+            page: other.page.as_deref().copied(),
+            hits_per_page: other.hits_per_page.as_deref().copied(),
+            attributes_to_retrieve: other.attributes_to_retrieve.map(|o| o.into_iter().collect()),
+            retrieve_vectors: other.retrieve_vectors.0,
+            attributes_to_crop: other.attributes_to_crop.map(|o| o.into_iter().collect()),
+            crop_length: other.crop_length.0,
+            attributes_to_highlight: other.attributes_to_highlight.map(|o| o.into_iter().collect()),
+            filter,
+            sort: other.sort.map(|attr| fix_sort_query_parameters(&attr)),
+            distinct: other.distinct,
+            show_matches_position: other.show_matches_position.0,
+            show_ranking_score: other.show_ranking_score.0,
+            show_ranking_score_details: other.show_ranking_score_details.0,
+            facets: other.facets.map(|o| o.into_iter().collect()),
+            highlight_pre_tag: other.highlight_pre_tag,
+            highlight_post_tag: other.highlight_post_tag,
+            crop_marker: other.crop_marker,
+            matching_strategy: other.matching_strategy,
+            attributes_to_search_on: other.attributes_to_search_on.map(|o| o.into_iter().collect()),
+            hybrid,
+            ranking_score_threshold: other.ranking_score_threshold.map(|o| o.0),
+            locales: other.locales.map(|o| o.into_iter().collect()),
+        })
+    }
+}
+
+// TODO: TAMO: split on :asc, and :desc, instead of doing some weird things
+
+/// Transform the sort query parameter into something that matches the post expected format.
+fn fix_sort_query_parameters(sort_query: &str) -> Vec<String> {
+    let mut sort_parameters = Vec::new();
+    let mut merge = false;
+    for current_sort in sort_query.trim_matches('"').split(',').map(|s| s.trim()) {
+        if current_sort.starts_with("_geoPoint(") {
+            sort_parameters.push(current_sort.to_string());
+            merge = true;
+        } else if merge && !sort_parameters.is_empty() {
+            let s = sort_parameters.last_mut().unwrap();
+            s.push(',');
+            s.push_str(current_sort);
+            if current_sort.ends_with("):desc") || current_sort.ends_with("):asc") {
+                merge = false;
+            }
+        } else {
+            sort_parameters.push(current_sort.to_string());
+            merge = false;
+        }
+    }
+    sort_parameters
+}
+
+pub async fn search_with_url_query(
+    index_scheduler: GuardedData<ActionPolicy<{ actions::SEARCH }>, Data<IndexScheduler>>,
+    search_queue: web::Data<SearchQueue>,
+    index_uid: web::Path<String>,
+    params: AwebQueryParameter<SearchQueryGet, DeserrQueryParamError>,
+    req: HttpRequest,
+    analytics: web::Data<Analytics>,
+) -> Result<HttpResponse, ResponseError> {
+    debug!(parameters = ?params, "Search get");
+    let index_uid = IndexUid::try_from(index_uid.into_inner())?;
+
+    let mut query: SearchQuery = params.into_inner().try_into()?;
+
+    // Tenant token search_rules.
+    if let Some(search_rules) = index_scheduler.filters().get_index_search_rules(&index_uid) {
+        add_search_rules(&mut query.filter, search_rules);
+    }
+
+    let mut aggregate = SearchAggregator::<SearchGET>::from_query(&query);
+
+    let index = index_scheduler.index(&index_uid)?;
+    let features = index_scheduler.features();
+
+    let search_kind = search_kind(&query, index_scheduler.get_ref(), &index, features)?;
+    let retrieve_vector = RetrieveVectors::new(query.retrieve_vectors, features)?;
+    let permit = search_queue.try_get_search_permit().await?;
+    let search_result = tokio::task::spawn_blocking(move || {
+        perform_search(&index, query, search_kind, retrieve_vector, index_scheduler.features())
+    })
+    .await;
+    permit.drop().await;
+    let search_result = search_result?;
+    if let Ok(ref search_result) = search_result {
+        aggregate.succeed(search_result);
+    }
+    analytics.publish(aggregate, &req);
+
+    let search_result = search_result?;
+
+    debug!(returns = ?search_result, "Search get");
+    Ok(HttpResponse::Ok().json(search_result))
+}
+
+pub async fn search_with_post(
+    index_scheduler: GuardedData<ActionPolicy<{ actions::SEARCH }>, Data<IndexScheduler>>,
+    search_queue: web::Data<SearchQueue>,
+    index_uid: web::Path<String>,
+    params: AwebJson<SearchQuery, DeserrJsonError>,
+    req: HttpRequest,
+    analytics: web::Data<Analytics>,
+) -> Result<HttpResponse, ResponseError> {
+    let index_uid = IndexUid::try_from(index_uid.into_inner())?;
+
+    let mut query = params.into_inner();
+    debug!(parameters = ?query, "Search post");
+
+    // Tenant token search_rules.
+    if let Some(search_rules) = index_scheduler.filters().get_index_search_rules(&index_uid) {
+        add_search_rules(&mut query.filter, search_rules);
+    }
+
+    let mut aggregate = SearchAggregator::<SearchPOST>::from_query(&query);
+
+    let index = index_scheduler.index(&index_uid)?;
+
+    let features = index_scheduler.features();
+
+    let search_kind = search_kind(&query, index_scheduler.get_ref(), &index, features)?;
+    let retrieve_vectors = RetrieveVectors::new(query.retrieve_vectors, features)?;
+
+    let permit = search_queue.try_get_search_permit().await?;
+    let search_result = tokio::task::spawn_blocking(move || {
+        perform_search(&index, query, search_kind, retrieve_vectors, index_scheduler.features())
+    })
+    .await;
+    permit.drop().await;
+    let search_result = search_result?;
+    if let Ok(ref search_result) = search_result {
+        aggregate.succeed(search_result);
+        if search_result.degraded {
+            MEILISEARCH_DEGRADED_SEARCH_REQUESTS.inc();
+        }
+    }
+    analytics.publish(aggregate, &req);
+
+    let search_result = search_result?;
+
+    debug!(returns = ?search_result, "Search post");
+    Ok(HttpResponse::Ok().json(search_result))
+}
+
+pub fn search_kind(
+    query: &SearchQuery,
+    index_scheduler: &IndexScheduler,
+    index: &milli::Index,
+    features: RoFeatures,
+) -> Result<SearchKind, ResponseError> {
+    if query.vector.is_some() {
+        features.check_vector("Passing `vector` as a parameter")?;
+    }
+    if query.hybrid.is_some() {
+        features.check_vector("Passing `hybrid` as a parameter")?;
+    }
+
+    // handle with care, the order of cases matters, the semantics is subtle
+    match (query.q.as_deref(), &query.hybrid, query.vector.as_deref()) {
+        // empty query, no vector => placeholder search
+        (Some(q), _, None) if q.trim().is_empty() => Ok(SearchKind::KeywordOnly),
+        // no query, no vector => placeholder search
+        (None, _, None) => Ok(SearchKind::KeywordOnly),
+        // hybrid.semantic_ratio == 1.0 => vector
+        (_, Some(HybridQuery { semantic_ratio, embedder }), v) if **semantic_ratio == 1.0 => {
+            SearchKind::semantic(index_scheduler, index, embedder, v.map(|v| v.len()))
+        }
+        // hybrid.semantic_ratio == 0.0 => keyword
+        (_, Some(HybridQuery { semantic_ratio, embedder: _ }), _) if **semantic_ratio == 0.0 => {
+            Ok(SearchKind::KeywordOnly)
+        }
+        // no query, hybrid, vector => semantic
+        (None, Some(HybridQuery { semantic_ratio: _, embedder }), Some(v)) => {
+            SearchKind::semantic(index_scheduler, index, embedder, Some(v.len()))
+        }
+        // query, no hybrid, no vector => keyword
+        (Some(_), None, None) => Ok(SearchKind::KeywordOnly),
+        // query, hybrid, maybe vector => hybrid
+        (Some(_), Some(HybridQuery { semantic_ratio, embedder }), v) => SearchKind::hybrid(
+            index_scheduler,
+            index,
+            embedder,
+            **semantic_ratio,
+            v.map(|v| v.len()),
+        ),
+
+        (_, None, Some(_)) => Err(MeilisearchHttpError::MissingSearchHybrid.into()),
+    }
+}
+
+#[cfg(test)]
+mod test {
+    use super::*;
+
+    #[test]
+    fn test_fix_sort_query_parameters() {
+        let sort = fix_sort_query_parameters("_geoPoint(12, 13):asc");
+        assert_eq!(sort, vec!["_geoPoint(12,13):asc".to_string()]);
+        let sort = fix_sort_query_parameters("doggo:asc,_geoPoint(12.45,13.56):desc");
+        assert_eq!(sort, vec!["doggo:asc".to_string(), "_geoPoint(12.45,13.56):desc".to_string(),]);
+        let sort = fix_sort_query_parameters(
+            "doggo:asc , _geoPoint(12.45, 13.56, 2590352):desc , catto:desc",
+        );
+        assert_eq!(
+            sort,
+            vec![
+                "doggo:asc".to_string(),
+                "_geoPoint(12.45,13.56,2590352):desc".to_string(),
+                "catto:desc".to_string(),
+            ]
+        );
+        let sort = fix_sort_query_parameters("doggo:asc , _geoPoint(1, 2), catto:desc");
+        // This is ugly but eh, I don't want to write a full parser just for this unused route
+        assert_eq!(sort, vec!["doggo:asc".to_string(), "_geoPoint(1,2),catto:desc".to_string(),]);
+    }
+}

crates/meilisearch/src/routes/indexes/search_analytics.rs

@@ -0,0 +1,485 @@
+use once_cell::sync::Lazy;
+use regex::Regex;
+use serde_json::{json, Value};
+use std::collections::{BTreeSet, BinaryHeap, HashMap};
+
+use meilisearch_types::locales::Locale;
+
+use crate::{
+    aggregate_methods,
+    analytics::{Aggregate, AggregateMethod},
+    search::{
+        SearchQuery, SearchResult, DEFAULT_CROP_LENGTH, DEFAULT_CROP_MARKER,
+        DEFAULT_HIGHLIGHT_POST_TAG, DEFAULT_HIGHLIGHT_PRE_TAG, DEFAULT_SEARCH_LIMIT,
+        DEFAULT_SEMANTIC_RATIO,
+    },
+};
+
+aggregate_methods!(
+    SearchGET => "Documents Searched GET",
+    SearchPOST => "Documents Searched POST",
+);
+
+#[derive(Default)]
+pub struct SearchAggregator<Method: AggregateMethod> {
+    // requests
+    total_received: usize,
+    total_succeeded: usize,
+    total_degraded: usize,
+    total_used_negative_operator: usize,
+    time_spent: BinaryHeap<usize>,
+
+    // sort
+    sort_with_geo_point: bool,
+    // every time a request has a filter, this field must be incremented by the number of terms it contains
+    sort_sum_of_criteria_terms: usize,
+    // every time a request has a filter, this field must be incremented by one
+    sort_total_number_of_criteria: usize,
+
+    // distinct
+    distinct: bool,
+
+    // filter
+    filter_with_geo_radius: bool,
+    filter_with_geo_bounding_box: bool,
+    // every time a request has a filter, this field must be incremented by the number of terms it contains
+    filter_sum_of_criteria_terms: usize,
+    // every time a request has a filter, this field must be incremented by one
+    filter_total_number_of_criteria: usize,
+    used_syntax: HashMap<String, usize>,
+
+    // attributes_to_search_on
+    // every time a search is done using attributes_to_search_on
+    attributes_to_search_on_total_number_of_uses: usize,
+
+    // q
+    // The maximum number of terms in a q request
+    max_terms_number: usize,
+
+    // vector
+    // The maximum number of floats in a vector request
+    max_vector_size: usize,
+    // Whether the semantic ratio passed to a hybrid search equals the default ratio.
+    semantic_ratio: bool,
+    hybrid: bool,
+    retrieve_vectors: bool,
+
+    // every time a search is done, we increment the counter linked to the used settings
+    matching_strategy: HashMap<String, usize>,
+
+    // List of the unique Locales passed as parameter
+    locales: BTreeSet<Locale>,
+
+    // pagination
+    max_limit: usize,
+    max_offset: usize,
+    finite_pagination: usize,
+
+    // formatting
+    max_attributes_to_retrieve: usize,
+    max_attributes_to_highlight: usize,
+    highlight_pre_tag: bool,
+    highlight_post_tag: bool,
+    max_attributes_to_crop: usize,
+    crop_marker: bool,
+    show_matches_position: bool,
+    crop_length: bool,
+
+    // facets
+    facets_sum_of_terms: usize,
+    facets_total_number_of_facets: usize,
+
+    // scoring
+    show_ranking_score: bool,
+    show_ranking_score_details: bool,
+    ranking_score_threshold: bool,
+
+    marker: std::marker::PhantomData<Method>,
+}
+
+impl<Method: AggregateMethod> SearchAggregator<Method> {
+    #[allow(clippy::field_reassign_with_default)]
+    pub fn from_query(query: &SearchQuery) -> Self {
+        let SearchQuery {
+            q,
+            vector,
+            offset,
+            limit,
+            page,
+            hits_per_page,
+            attributes_to_retrieve: _,
+            retrieve_vectors,
+            attributes_to_crop: _,
+            crop_length,
+            attributes_to_highlight: _,
+            show_matches_position,
+            show_ranking_score,
+            show_ranking_score_details,
+            filter,
+            sort,
+            distinct,
+            facets: _,
+            highlight_pre_tag,
+            highlight_post_tag,
+            crop_marker,
+            matching_strategy,
+            attributes_to_search_on,
+            hybrid,
+            ranking_score_threshold,
+            locales,
+        } = query;
+
+        let mut ret = Self::default();
+
+        ret.total_received = 1;
+
+        if let Some(ref sort) = sort {
+            ret.sort_total_number_of_criteria = 1;
+            ret.sort_with_geo_point = sort.iter().any(|s| s.contains("_geoPoint("));
+            ret.sort_sum_of_criteria_terms = sort.len();
+        }
+
+        ret.distinct = distinct.is_some();
+
+        if let Some(ref filter) = filter {
+            static RE: Lazy<Regex> = Lazy::new(|| Regex::new("AND | OR").unwrap());
+            ret.filter_total_number_of_criteria = 1;
+
+            let syntax = match filter {
+                Value::String(_) => "string".to_string(),
+                Value::Array(values) => {
+                    if values.iter().map(|v| v.to_string()).any(|s| RE.is_match(&s)) {
+                        "mixed".to_string()
+                    } else {
+                        "array".to_string()
+                    }
+                }
+                _ => "none".to_string(),
+            };
+            // convert the string to a HashMap
+            ret.used_syntax.insert(syntax, 1);
+
+            let stringified_filters = filter.to_string();
+            ret.filter_with_geo_radius = stringified_filters.contains("_geoRadius(");
+            ret.filter_with_geo_bounding_box = stringified_filters.contains("_geoBoundingBox(");
+            ret.filter_sum_of_criteria_terms = RE.split(&stringified_filters).count();
+        }
+
+        // attributes_to_search_on
+        if attributes_to_search_on.is_some() {
+            ret.attributes_to_search_on_total_number_of_uses = 1;
+        }
+
+        if let Some(ref q) = q {
+            ret.max_terms_number = q.split_whitespace().count();
+        }
+
+        if let Some(ref vector) = vector {
+            ret.max_vector_size = vector.len();
+        }
+        ret.retrieve_vectors |= retrieve_vectors;
+
+        if query.is_finite_pagination() {
+            let limit = hits_per_page.unwrap_or_else(DEFAULT_SEARCH_LIMIT);
+            ret.max_limit = limit;
+            ret.max_offset = page.unwrap_or(1).saturating_sub(1) * limit;
+            ret.finite_pagination = 1;
+        } else {
+            ret.max_limit = *limit;
+            ret.max_offset = *offset;
+            ret.finite_pagination = 0;
+        }
+
+        ret.matching_strategy.insert(format!("{:?}", matching_strategy), 1);
+
+        if let Some(locales) = locales {
+            ret.locales = locales.iter().copied().collect();
+        }
+
+        ret.highlight_pre_tag = *highlight_pre_tag != DEFAULT_HIGHLIGHT_PRE_TAG();
+        ret.highlight_post_tag = *highlight_post_tag != DEFAULT_HIGHLIGHT_POST_TAG();
+        ret.crop_marker = *crop_marker != DEFAULT_CROP_MARKER();
+        ret.crop_length = *crop_length != DEFAULT_CROP_LENGTH();
+        ret.show_matches_position = *show_matches_position;
+
+        ret.show_ranking_score = *show_ranking_score;
+        ret.show_ranking_score_details = *show_ranking_score_details;
+        ret.ranking_score_threshold = ranking_score_threshold.is_some();
+
+        if let Some(hybrid) = hybrid {
+            ret.semantic_ratio = hybrid.semantic_ratio != DEFAULT_SEMANTIC_RATIO();
+            ret.hybrid = true;
+        }
+
+        ret
+    }
+
+    pub fn succeed(&mut self, result: &SearchResult) {
+        let SearchResult {
+            hits: _,
+            query: _,
+            processing_time_ms,
+            hits_info: _,
+            semantic_hit_count: _,
+            facet_distribution: _,
+            facet_stats: _,
+            degraded,
+            used_negative_operator,
+        } = result;
+
+        self.total_succeeded = self.total_succeeded.saturating_add(1);
+        if *degraded {
+            self.total_degraded = self.total_degraded.saturating_add(1);
+        }
+        if *used_negative_operator {
+            self.total_used_negative_operator = self.total_used_negative_operator.saturating_add(1);
+        }
+        self.time_spent.push(*processing_time_ms as usize);
+    }
+}
+
+impl<Method: AggregateMethod> Aggregate for SearchAggregator<Method> {
+    fn event_name(&self) -> &'static str {
+        Method::event_name()
+    }
+
+    fn aggregate(mut self: Box<Self>, new: Box<Self>) -> Box<Self> {
+        let Self {
+            total_received,
+            total_succeeded,
+            mut time_spent,
+            sort_with_geo_point,
+            sort_sum_of_criteria_terms,
+            sort_total_number_of_criteria,
+            distinct,
+            filter_with_geo_radius,
+            filter_with_geo_bounding_box,
+            filter_sum_of_criteria_terms,
+            filter_total_number_of_criteria,
+            used_syntax,
+            attributes_to_search_on_total_number_of_uses,
+            max_terms_number,
+            max_vector_size,
+            retrieve_vectors,
+            matching_strategy,
+            max_limit,
+            max_offset,
+            finite_pagination,
+            max_attributes_to_retrieve,
+            max_attributes_to_highlight,
+            highlight_pre_tag,
+            highlight_post_tag,
+            max_attributes_to_crop,
+            crop_marker,
+            show_matches_position,
+            crop_length,
+            facets_sum_of_terms,
+            facets_total_number_of_facets,
+            show_ranking_score,
+            show_ranking_score_details,
+            semantic_ratio,
+            hybrid,
+            total_degraded,
+            total_used_negative_operator,
+            ranking_score_threshold,
+            mut locales,
+            marker: _,
+        } = *new;
+
+        // request
+        self.total_received = self.total_received.saturating_add(total_received);
+        self.total_succeeded = self.total_succeeded.saturating_add(total_succeeded);
+        self.total_degraded = self.total_degraded.saturating_add(total_degraded);
+        self.total_used_negative_operator =
+            self.total_used_negative_operator.saturating_add(total_used_negative_operator);
+        self.time_spent.append(&mut time_spent);
+
+        // sort
+        self.sort_with_geo_point |= sort_with_geo_point;
+        self.sort_sum_of_criteria_terms =
+            self.sort_sum_of_criteria_terms.saturating_add(sort_sum_of_criteria_terms);
+        self.sort_total_number_of_criteria =
+            self.sort_total_number_of_criteria.saturating_add(sort_total_number_of_criteria);
+
+        // distinct
+        self.distinct |= distinct;
+
+        // filter
+        self.filter_with_geo_radius |= filter_with_geo_radius;
+        self.filter_with_geo_bounding_box |= filter_with_geo_bounding_box;
+        self.filter_sum_of_criteria_terms =
+            self.filter_sum_of_criteria_terms.saturating_add(filter_sum_of_criteria_terms);
+        self.filter_total_number_of_criteria =
+            self.filter_total_number_of_criteria.saturating_add(filter_total_number_of_criteria);
+        for (key, value) in used_syntax.into_iter() {
+            let used_syntax = self.used_syntax.entry(key).or_insert(0);
+            *used_syntax = used_syntax.saturating_add(value);
+        }
+
+        // attributes_to_search_on
+        self.attributes_to_search_on_total_number_of_uses = self
+            .attributes_to_search_on_total_number_of_uses
+            .saturating_add(attributes_to_search_on_total_number_of_uses);
+
+        // q
+        self.max_terms_number = self.max_terms_number.max(max_terms_number);
+
+        // vector
+        self.max_vector_size = self.max_vector_size.max(max_vector_size);
+        self.retrieve_vectors |= retrieve_vectors;
+        self.semantic_ratio |= semantic_ratio;
+        self.hybrid |= hybrid;
+
+        // pagination
+        self.max_limit = self.max_limit.max(max_limit);
+        self.max_offset = self.max_offset.max(max_offset);
+        self.finite_pagination += finite_pagination;
+
+        // formatting
+        self.max_attributes_to_retrieve =
+            self.max_attributes_to_retrieve.max(max_attributes_to_retrieve);
+        self.max_attributes_to_highlight =
+            self.max_attributes_to_highlight.max(max_attributes_to_highlight);
+        self.highlight_pre_tag |= highlight_pre_tag;
+        self.highlight_post_tag |= highlight_post_tag;
+        self.max_attributes_to_crop = self.max_attributes_to_crop.max(max_attributes_to_crop);
+        self.crop_marker |= crop_marker;
+        self.show_matches_position |= show_matches_position;
+        self.crop_length |= crop_length;
+
+        // facets
+        self.facets_sum_of_terms = self.facets_sum_of_terms.saturating_add(facets_sum_of_terms);
+        self.facets_total_number_of_facets =
+            self.facets_total_number_of_facets.saturating_add(facets_total_number_of_facets);
+
+        // matching strategy
+        for (key, value) in matching_strategy.into_iter() {
+            let matching_strategy = self.matching_strategy.entry(key).or_insert(0);
+            *matching_strategy = matching_strategy.saturating_add(value);
+        }
+
+        // scoring
+        self.show_ranking_score |= show_ranking_score;
+        self.show_ranking_score_details |= show_ranking_score_details;
+        self.ranking_score_threshold |= ranking_score_threshold;
+
+        // locales
+        self.locales.append(&mut locales);
+
+        self
+    }
+
+    fn into_event(self: Box<Self>) -> serde_json::Value {
+        let Self {
+            total_received,
+            total_succeeded,
+            time_spent,
+            sort_with_geo_point,
+            sort_sum_of_criteria_terms,
+            sort_total_number_of_criteria,
+            distinct,
+            filter_with_geo_radius,
+            filter_with_geo_bounding_box,
+            filter_sum_of_criteria_terms,
+            filter_total_number_of_criteria,
+            used_syntax,
+            attributes_to_search_on_total_number_of_uses,
+            max_terms_number,
+            max_vector_size,
+            retrieve_vectors,
+            matching_strategy,
+            max_limit,
+            max_offset,
+            finite_pagination,
+            max_attributes_to_retrieve,
+            max_attributes_to_highlight,
+            highlight_pre_tag,
+            highlight_post_tag,
+            max_attributes_to_crop,
+            crop_marker,
+            show_matches_position,
+            crop_length,
+            facets_sum_of_terms,
+            facets_total_number_of_facets,
+            show_ranking_score,
+            show_ranking_score_details,
+            semantic_ratio,
+            hybrid,
+            total_degraded,
+            total_used_negative_operator,
+            ranking_score_threshold,
+            locales,
+            marker: _,
+        } = *self;
+
+        // we get all the values in a sorted manner
+        let time_spent = time_spent.into_sorted_vec();
+        // the index of the 99th percentage of value
+        let percentile_99th = time_spent.len() * 99 / 100;
+        // We are only interested by the slowest value of the 99th fastest results
+        let time_spent = time_spent.get(percentile_99th);
+
+        json!({
+            "requests": {
+                "99th_response_time": time_spent.map(|t| format!("{:.2}", t)),
+                "total_succeeded": total_succeeded,
+                "total_failed": total_received.saturating_sub(total_succeeded), // just to be sure we never panics
+                "total_received": total_received,
+                "total_degraded": total_degraded,
+                "total_used_negative_operator": total_used_negative_operator,
+            },
+            "sort": {
+                "with_geoPoint": sort_with_geo_point,
+                "avg_criteria_number": format!("{:.2}", sort_sum_of_criteria_terms as f64 / sort_total_number_of_criteria as f64),
+            },
+            "distinct": distinct,
+            "filter": {
+               "with_geoRadius": filter_with_geo_radius,
+               "with_geoBoundingBox": filter_with_geo_bounding_box,
+               "avg_criteria_number": format!("{:.2}", filter_sum_of_criteria_terms as f64 / filter_total_number_of_criteria as f64),
+               "most_used_syntax": used_syntax.iter().max_by_key(|(_, v)| *v).map(|(k, _)| json!(k)).unwrap_or_else(|| json!(null)),
+            },
+            "attributes_to_search_on": {
+               "total_number_of_uses": attributes_to_search_on_total_number_of_uses,
+            },
+            "q": {
+               "max_terms_number": max_terms_number,
+            },
+            "vector": {
+                "max_vector_size": max_vector_size,
+                "retrieve_vectors": retrieve_vectors,
+            },
+            "hybrid": {
+                "enabled": hybrid,
+                "semantic_ratio": semantic_ratio,
+            },
+            "pagination": {
+               "max_limit": max_limit,
+               "max_offset": max_offset,
+               "most_used_navigation": if finite_pagination > (total_received / 2) { "exhaustive" } else { "estimated" },
+            },
+            "formatting": {
+                "max_attributes_to_retrieve": max_attributes_to_retrieve,
+                "max_attributes_to_highlight": max_attributes_to_highlight,
+                "highlight_pre_tag": highlight_pre_tag,
+                "highlight_post_tag": highlight_post_tag,
+                "max_attributes_to_crop": max_attributes_to_crop,
+                "crop_marker": crop_marker,
+                "show_matches_position": show_matches_position,
+                "crop_length": crop_length,
+            },
+            "facets": {
+                "avg_facets_number": format!("{:.2}", facets_sum_of_terms as f64 / facets_total_number_of_facets as f64),
+            },
+            "matching_strategy": {
+                "most_used_strategy": matching_strategy.iter().max_by_key(|(_, v)| *v).map(|(k, _)| json!(k)).unwrap_or_else(|| json!(null)),
+            },
+            "locales": locales,
+            "scoring": {
+                "show_ranking_score": show_ranking_score,
+                "show_ranking_score_details": show_ranking_score_details,
+                "ranking_score_threshold": ranking_score_threshold,
+            },
+        })
+    }
+}

crates/meilisearch/src/routes/indexes/settings.rs

@@ -0,0 +1,532 @@
+use super::settings_analytics::*;
+use actix_web::web::Data;
+use actix_web::{web, HttpRequest, HttpResponse};
+use deserr::actix_web::AwebJson;
+use index_scheduler::IndexScheduler;
+use meilisearch_types::deserr::DeserrJsonError;
+use meilisearch_types::error::ResponseError;
+use meilisearch_types::index_uid::IndexUid;
+use meilisearch_types::milli::update::Setting;
+use meilisearch_types::settings::{settings, SecretPolicy, Settings, Unchecked};
+use meilisearch_types::tasks::KindWithContent;
+use tracing::debug;
+
+use crate::analytics::Analytics;
+use crate::extractors::authentication::policies::*;
+use crate::extractors::authentication::GuardedData;
+use crate::routes::{get_task_id, is_dry_run, SummarizedTaskView};
+use crate::Opt;
+
+#[macro_export]
+macro_rules! make_setting_route {
+    ($route:literal, $update_verb:ident, $type:ty, $err_ty:ty, $attr:ident, $camelcase_attr:literal, $analytics:ident) => {
+        pub mod $attr {
+            use actix_web::web::Data;
+            use actix_web::{web, HttpRequest, HttpResponse, Resource};
+            use index_scheduler::IndexScheduler;
+            use meilisearch_types::error::ResponseError;
+            use meilisearch_types::index_uid::IndexUid;
+            use meilisearch_types::milli::update::Setting;
+            use meilisearch_types::settings::{settings, Settings};
+            use meilisearch_types::tasks::KindWithContent;
+            use tracing::debug;
+            use $crate::analytics::Analytics;
+            use $crate::extractors::authentication::policies::*;
+            use $crate::extractors::authentication::GuardedData;
+            use $crate::extractors::sequential_extractor::SeqHandler;
+            use $crate::Opt;
+            use $crate::routes::{is_dry_run, get_task_id, SummarizedTaskView};
+
+            pub async fn delete(
+                index_scheduler: GuardedData<
+                    ActionPolicy<{ actions::SETTINGS_UPDATE }>,
+                    Data<IndexScheduler>,
+                >,
+                index_uid: web::Path<String>,
+                req: HttpRequest,
+                opt: web::Data<Opt>,
+            ) -> Result<HttpResponse, ResponseError> {
+                let index_uid = IndexUid::try_from(index_uid.into_inner())?;
+
+                let new_settings = Settings { $attr: Setting::Reset.into(), ..Default::default() };
+
+                let allow_index_creation =
+                    index_scheduler.filters().allow_index_creation(&index_uid);
+
+                let task = KindWithContent::SettingsUpdate {
+                    index_uid: index_uid.to_string(),
+                    new_settings: Box::new(new_settings),
+                    is_deletion: true,
+                    allow_index_creation,
+                };
+                let uid = get_task_id(&req, &opt)?;
+                let dry_run = is_dry_run(&req, &opt)?;
+                let task: SummarizedTaskView =
+                    tokio::task::spawn_blocking(move || index_scheduler.register(task, uid, dry_run))
+                        .await??
+                        .into();
+
+                debug!(returns = ?task, "Delete settings");
+                Ok(HttpResponse::Accepted().json(task))
+            }
+
+            pub async fn update(
+                index_scheduler: GuardedData<
+                    ActionPolicy<{ actions::SETTINGS_UPDATE }>,
+                    Data<IndexScheduler>,
+                >,
+                index_uid: actix_web::web::Path<String>,
+                body: deserr::actix_web::AwebJson<Option<$type>, $err_ty>,
+                req: HttpRequest,
+                opt: web::Data<Opt>,
+                analytics: web::Data<Analytics>,
+            ) -> std::result::Result<HttpResponse, ResponseError> {
+                let index_uid = IndexUid::try_from(index_uid.into_inner())?;
+
+                let body = body.into_inner();
+                debug!(parameters = ?body, "Update settings");
+
+                #[allow(clippy::redundant_closure_call)]
+                analytics.publish(
+                    $crate::routes::indexes::settings_analytics::$analytics::new(body.as_ref()).into_settings(),
+                    &req,
+                );
+
+                let new_settings = Settings {
+                    $attr: match body {
+                        Some(inner_body) => Setting::Set(inner_body).into(),
+                        None => Setting::Reset.into(),
+                    },
+                    ..Default::default()
+                };
+
+                let new_settings = $crate::routes::indexes::settings::validate_settings(
+                    new_settings,
+                    &index_scheduler,
+                )?;
+
+                let allow_index_creation =
+                    index_scheduler.filters().allow_index_creation(&index_uid);
+
+                let task = KindWithContent::SettingsUpdate {
+                    index_uid: index_uid.to_string(),
+                    new_settings: Box::new(new_settings),
+                    is_deletion: false,
+                    allow_index_creation,
+                };
+                let uid = get_task_id(&req, &opt)?;
+                let dry_run = is_dry_run(&req, &opt)?;
+                let task: SummarizedTaskView =
+                    tokio::task::spawn_blocking(move || index_scheduler.register(task, uid, dry_run))
+                        .await??
+                        .into();
+
+                debug!(returns = ?task, "Update settings");
+                Ok(HttpResponse::Accepted().json(task))
+            }
+
+            pub async fn get(
+                index_scheduler: GuardedData<
+                    ActionPolicy<{ actions::SETTINGS_GET }>,
+                    Data<IndexScheduler>,
+                >,
+                index_uid: actix_web::web::Path<String>,
+            ) -> std::result::Result<HttpResponse, ResponseError> {
+                let index_uid = IndexUid::try_from(index_uid.into_inner())?;
+
+                let index = index_scheduler.index(&index_uid)?;
+                let rtxn = index.read_txn()?;
+                let settings = settings(&index, &rtxn, meilisearch_types::settings::SecretPolicy::HideSecrets)?;
+
+                debug!(returns = ?settings, "Update settings");
+
+                Ok(HttpResponse::Ok().json(settings.$attr))
+            }
+
+            pub fn resources() -> Resource {
+                Resource::new($route)
+                    .route(web::get().to(SeqHandler(get)))
+                    .route(web::$update_verb().to(SeqHandler(update)))
+                    .route(web::delete().to(SeqHandler(delete)))
+            }
+        }
+    };
+}
+
+make_setting_route!(
+    "/filterable-attributes",
+    put,
+    std::collections::BTreeSet<String>,
+    meilisearch_types::deserr::DeserrJsonError<
+        meilisearch_types::error::deserr_codes::InvalidSettingsFilterableAttributes,
+    >,
+    filterable_attributes,
+    "filterableAttributes",
+    FilterableAttributesAnalytics
+);
+
+make_setting_route!(
+    "/sortable-attributes",
+    put,
+    std::collections::BTreeSet<String>,
+    meilisearch_types::deserr::DeserrJsonError<
+        meilisearch_types::error::deserr_codes::InvalidSettingsSortableAttributes,
+    >,
+    sortable_attributes,
+    "sortableAttributes",
+    SortableAttributesAnalytics
+);
+
+make_setting_route!(
+    "/displayed-attributes",
+    put,
+    Vec<String>,
+    meilisearch_types::deserr::DeserrJsonError<
+        meilisearch_types::error::deserr_codes::InvalidSettingsDisplayedAttributes,
+    >,
+    displayed_attributes,
+    "displayedAttributes",
+    DisplayedAttributesAnalytics
+);
+
+make_setting_route!(
+    "/typo-tolerance",
+    patch,
+    meilisearch_types::settings::TypoSettings,
+    meilisearch_types::deserr::DeserrJsonError<
+        meilisearch_types::error::deserr_codes::InvalidSettingsTypoTolerance,
+    >,
+    typo_tolerance,
+    "typoTolerance",
+    TypoToleranceAnalytics
+);
+
+make_setting_route!(
+    "/searchable-attributes",
+    put,
+    Vec<String>,
+    meilisearch_types::deserr::DeserrJsonError<
+        meilisearch_types::error::deserr_codes::InvalidSettingsSearchableAttributes,
+    >,
+    searchable_attributes,
+    "searchableAttributes",
+    SearchableAttributesAnalytics
+);
+
+make_setting_route!(
+    "/stop-words",
+    put,
+    std::collections::BTreeSet<String>,
+    meilisearch_types::deserr::DeserrJsonError<
+        meilisearch_types::error::deserr_codes::InvalidSettingsStopWords,
+    >,
+    stop_words,
+    "stopWords",
+    StopWordsAnalytics
+);
+
+make_setting_route!(
+    "/non-separator-tokens",
+    put,
+    std::collections::BTreeSet<String>,
+    meilisearch_types::deserr::DeserrJsonError<
+        meilisearch_types::error::deserr_codes::InvalidSettingsNonSeparatorTokens,
+    >,
+    non_separator_tokens,
+    "nonSeparatorTokens",
+    NonSeparatorTokensAnalytics
+);
+
+make_setting_route!(
+    "/separator-tokens",
+    put,
+    std::collections::BTreeSet<String>,
+    meilisearch_types::deserr::DeserrJsonError<
+        meilisearch_types::error::deserr_codes::InvalidSettingsSeparatorTokens,
+    >,
+    separator_tokens,
+    "separatorTokens",
+    SeparatorTokensAnalytics
+);
+
+make_setting_route!(
+    "/dictionary",
+    put,
+    std::collections::BTreeSet<String>,
+    meilisearch_types::deserr::DeserrJsonError<
+        meilisearch_types::error::deserr_codes::InvalidSettingsDictionary,
+    >,
+    dictionary,
+    "dictionary",
+    DictionaryAnalytics
+);
+
+make_setting_route!(
+    "/synonyms",
+    put,
+    std::collections::BTreeMap<String, Vec<String>>,
+    meilisearch_types::deserr::DeserrJsonError<
+        meilisearch_types::error::deserr_codes::InvalidSettingsSynonyms,
+    >,
+    synonyms,
+    "synonyms",
+    SynonymsAnalytics
+);
+
+make_setting_route!(
+    "/distinct-attribute",
+    put,
+    String,
+    meilisearch_types::deserr::DeserrJsonError<
+        meilisearch_types::error::deserr_codes::InvalidSettingsDistinctAttribute,
+    >,
+    distinct_attribute,
+    "distinctAttribute",
+    DistinctAttributeAnalytics
+);
+
+make_setting_route!(
+    "/proximity-precision",
+    put,
+    meilisearch_types::settings::ProximityPrecisionView,
+    meilisearch_types::deserr::DeserrJsonError<
+        meilisearch_types::error::deserr_codes::InvalidSettingsProximityPrecision,
+    >,
+    proximity_precision,
+    "proximityPrecision",
+    ProximityPrecisionAnalytics
+);
+
+make_setting_route!(
+    "/localized-attributes",
+    put,
+    Vec<meilisearch_types::locales::LocalizedAttributesRuleView>,
+    meilisearch_types::deserr::DeserrJsonError<
+        meilisearch_types::error::deserr_codes::InvalidSettingsLocalizedAttributes,
+    >,
+    localized_attributes,
+    "localizedAttributes",
+    LocalesAnalytics
+);
+
+make_setting_route!(
+    "/ranking-rules",
+    put,
+    Vec<meilisearch_types::settings::RankingRuleView>,
+    meilisearch_types::deserr::DeserrJsonError<
+        meilisearch_types::error::deserr_codes::InvalidSettingsRankingRules,
+    >,
+    ranking_rules,
+    "rankingRules",
+    RankingRulesAnalytics
+);
+
+make_setting_route!(
+    "/faceting",
+    patch,
+    meilisearch_types::settings::FacetingSettings,
+    meilisearch_types::deserr::DeserrJsonError<
+        meilisearch_types::error::deserr_codes::InvalidSettingsFaceting,
+    >,
+    faceting,
+    "faceting",
+    FacetingAnalytics
+);
+
+make_setting_route!(
+    "/pagination",
+    patch,
+    meilisearch_types::settings::PaginationSettings,
+    meilisearch_types::deserr::DeserrJsonError<
+        meilisearch_types::error::deserr_codes::InvalidSettingsPagination,
+    >,
+    pagination,
+    "pagination",
+    PaginationAnalytics
+);
+
+make_setting_route!(
+    "/embedders",
+    patch,
+    std::collections::BTreeMap<String, Setting<meilisearch_types::milli::vector::settings::EmbeddingSettings>>,
+    meilisearch_types::deserr::DeserrJsonError<
+        meilisearch_types::error::deserr_codes::InvalidSettingsEmbedders,
+    >,
+    embedders,
+    "embedders",
+    EmbeddersAnalytics
+);
+
+make_setting_route!(
+    "/search-cutoff-ms",
+    put,
+    u64,
+    meilisearch_types::deserr::DeserrJsonError<
+        meilisearch_types::error::deserr_codes::InvalidSettingsSearchCutoffMs,
+    >,
+    search_cutoff_ms,
+    "searchCutoffMs",
+    SearchCutoffMsAnalytics
+);
+
+macro_rules! generate_configure {
+    ($($mod:ident),*) => {
+        pub fn configure(cfg: &mut web::ServiceConfig) {
+            use crate::extractors::sequential_extractor::SeqHandler;
+            cfg.service(
+                web::resource("")
+                .route(web::patch().to(SeqHandler(update_all)))
+                .route(web::get().to(SeqHandler(get_all)))
+                .route(web::delete().to(SeqHandler(delete_all))))
+                $(.service($mod::resources()))*;
+        }
+    };
+}
+
+generate_configure!(
+    filterable_attributes,
+    sortable_attributes,
+    displayed_attributes,
+    localized_attributes,
+    searchable_attributes,
+    distinct_attribute,
+    proximity_precision,
+    stop_words,
+    separator_tokens,
+    non_separator_tokens,
+    dictionary,
+    synonyms,
+    ranking_rules,
+    typo_tolerance,
+    pagination,
+    faceting,
+    embedders,
+    search_cutoff_ms
+);
+
+pub async fn update_all(
+    index_scheduler: GuardedData<ActionPolicy<{ actions::SETTINGS_UPDATE }>, Data<IndexScheduler>>,
+    index_uid: web::Path<String>,
+    body: AwebJson<Settings<Unchecked>, DeserrJsonError>,
+    req: HttpRequest,
+    opt: web::Data<Opt>,
+    analytics: web::Data<Analytics>,
+) -> Result<HttpResponse, ResponseError> {
+    let index_uid = IndexUid::try_from(index_uid.into_inner())?;
+
+    let new_settings = body.into_inner();
+    debug!(parameters = ?new_settings, "Update all settings");
+    let new_settings = validate_settings(new_settings, &index_scheduler)?;
+
+    analytics.publish(
+        SettingsAnalytics {
+            ranking_rules: RankingRulesAnalytics::new(new_settings.ranking_rules.as_ref().set()),
+            searchable_attributes: SearchableAttributesAnalytics::new(
+                new_settings.searchable_attributes.as_ref().set(),
+            ),
+            displayed_attributes: DisplayedAttributesAnalytics::new(
+                new_settings.displayed_attributes.as_ref().set(),
+            ),
+            sortable_attributes: SortableAttributesAnalytics::new(
+                new_settings.sortable_attributes.as_ref().set(),
+            ),
+            filterable_attributes: FilterableAttributesAnalytics::new(
+                new_settings.filterable_attributes.as_ref().set(),
+            ),
+            distinct_attribute: DistinctAttributeAnalytics::new(
+                new_settings.distinct_attribute.as_ref().set(),
+            ),
+            proximity_precision: ProximityPrecisionAnalytics::new(
+                new_settings.proximity_precision.as_ref().set(),
+            ),
+            typo_tolerance: TypoToleranceAnalytics::new(new_settings.typo_tolerance.as_ref().set()),
+            faceting: FacetingAnalytics::new(new_settings.faceting.as_ref().set()),
+            pagination: PaginationAnalytics::new(new_settings.pagination.as_ref().set()),
+            stop_words: StopWordsAnalytics::new(new_settings.stop_words.as_ref().set()),
+            synonyms: SynonymsAnalytics::new(new_settings.synonyms.as_ref().set()),
+            embedders: EmbeddersAnalytics::new(new_settings.embedders.as_ref().set()),
+            search_cutoff_ms: SearchCutoffMsAnalytics::new(
+                new_settings.search_cutoff_ms.as_ref().set(),
+            ),
+            locales: LocalesAnalytics::new(new_settings.localized_attributes.as_ref().set()),
+            dictionary: DictionaryAnalytics::new(new_settings.dictionary.as_ref().set()),
+            separator_tokens: SeparatorTokensAnalytics::new(
+                new_settings.separator_tokens.as_ref().set(),
+            ),
+            non_separator_tokens: NonSeparatorTokensAnalytics::new(
+                new_settings.non_separator_tokens.as_ref().set(),
+            ),
+        },
+        &req,
+    );
+
+    let allow_index_creation = index_scheduler.filters().allow_index_creation(&index_uid);
+    let index_uid = IndexUid::try_from(index_uid.into_inner())?.into_inner();
+    let task = KindWithContent::SettingsUpdate {
+        index_uid,
+        new_settings: Box::new(new_settings),
+        is_deletion: false,
+        allow_index_creation,
+    };
+    let uid = get_task_id(&req, &opt)?;
+    let dry_run = is_dry_run(&req, &opt)?;
+    let task: SummarizedTaskView =
+        tokio::task::spawn_blocking(move || index_scheduler.register(task, uid, dry_run))
+            .await??
+            .into();
+
+    debug!(returns = ?task, "Update all settings");
+    Ok(HttpResponse::Accepted().json(task))
+}
+
+pub async fn get_all(
+    index_scheduler: GuardedData<ActionPolicy<{ actions::SETTINGS_GET }>, Data<IndexScheduler>>,
+    index_uid: web::Path<String>,
+) -> Result<HttpResponse, ResponseError> {
+    let index_uid = IndexUid::try_from(index_uid.into_inner())?;
+
+    let index = index_scheduler.index(&index_uid)?;
+    let rtxn = index.read_txn()?;
+    let new_settings = settings(&index, &rtxn, SecretPolicy::HideSecrets)?;
+    debug!(returns = ?new_settings, "Get all settings");
+    Ok(HttpResponse::Ok().json(new_settings))
+}
+
+pub async fn delete_all(
+    index_scheduler: GuardedData<ActionPolicy<{ actions::SETTINGS_UPDATE }>, Data<IndexScheduler>>,
+    index_uid: web::Path<String>,
+    req: HttpRequest,
+    opt: web::Data<Opt>,
+) -> Result<HttpResponse, ResponseError> {
+    let index_uid = IndexUid::try_from(index_uid.into_inner())?;
+
+    let new_settings = Settings::cleared().into_unchecked();
+
+    let allow_index_creation = index_scheduler.filters().allow_index_creation(&index_uid);
+    let index_uid = IndexUid::try_from(index_uid.into_inner())?.into_inner();
+    let task = KindWithContent::SettingsUpdate {
+        index_uid,
+        new_settings: Box::new(new_settings),
+        is_deletion: true,
+        allow_index_creation,
+    };
+    let uid = get_task_id(&req, &opt)?;
+    let dry_run = is_dry_run(&req, &opt)?;
+    let task: SummarizedTaskView =
+        tokio::task::spawn_blocking(move || index_scheduler.register(task, uid, dry_run))
+            .await??
+            .into();
+
+    debug!(returns = ?task, "Delete all settings");
+    Ok(HttpResponse::Accepted().json(task))
+}
+
+fn validate_settings(
+    settings: Settings<Unchecked>,
+    index_scheduler: &IndexScheduler,
+) -> Result<Settings<Unchecked>, ResponseError> {
+    if matches!(settings.embedders, Setting::Set(_)) {
+        index_scheduler.features().check_vector("Passing `embedders` in settings")?
+    }
+    Ok(settings.validate()?)
+}

crates/meilisearch/src/routes/indexes/settings_analytics.rs

@@ -0,0 +1,621 @@
+//! All the structures used to make the analytics on the settings works.
+//! The signatures of the `new` functions are not very rust idiomatic because they must match the types received
+//! through the sub-settings route directly without any manipulation.
+//! This is why we often use a `Option<&Vec<_>>` instead of a `Option<&[_]>`.
+
+use meilisearch_types::locales::{Locale, LocalizedAttributesRuleView};
+use meilisearch_types::milli::update::Setting;
+use meilisearch_types::milli::vector::settings::EmbeddingSettings;
+use meilisearch_types::settings::{
+    FacetingSettings, PaginationSettings, ProximityPrecisionView, TypoSettings,
+};
+use meilisearch_types::{facet_values_sort::FacetValuesSort, settings::RankingRuleView};
+use serde::Serialize;
+use std::collections::{BTreeMap, BTreeSet, HashSet};
+
+use crate::analytics::Aggregate;
+
+#[derive(Serialize, Default)]
+pub struct SettingsAnalytics {
+    pub ranking_rules: RankingRulesAnalytics,
+    pub searchable_attributes: SearchableAttributesAnalytics,
+    pub displayed_attributes: DisplayedAttributesAnalytics,
+    pub sortable_attributes: SortableAttributesAnalytics,
+    pub filterable_attributes: FilterableAttributesAnalytics,
+    pub distinct_attribute: DistinctAttributeAnalytics,
+    pub proximity_precision: ProximityPrecisionAnalytics,
+    pub typo_tolerance: TypoToleranceAnalytics,
+    pub faceting: FacetingAnalytics,
+    pub pagination: PaginationAnalytics,
+    pub stop_words: StopWordsAnalytics,
+    pub synonyms: SynonymsAnalytics,
+    pub embedders: EmbeddersAnalytics,
+    pub search_cutoff_ms: SearchCutoffMsAnalytics,
+    pub locales: LocalesAnalytics,
+    pub dictionary: DictionaryAnalytics,
+    pub separator_tokens: SeparatorTokensAnalytics,
+    pub non_separator_tokens: NonSeparatorTokensAnalytics,
+}
+
+impl Aggregate for SettingsAnalytics {
+    fn event_name(&self) -> &'static str {
+        "Settings Updated"
+    }
+
+    fn aggregate(self: Box<Self>, new: Box<Self>) -> Box<Self> {
+        Box::new(Self {
+            ranking_rules: RankingRulesAnalytics {
+                words_position: new
+                    .ranking_rules
+                    .words_position
+                    .or(self.ranking_rules.words_position),
+                typo_position: new.ranking_rules.typo_position.or(self.ranking_rules.typo_position),
+                proximity_position: new
+                    .ranking_rules
+                    .proximity_position
+                    .or(self.ranking_rules.proximity_position),
+                attribute_position: new
+                    .ranking_rules
+                    .attribute_position
+                    .or(self.ranking_rules.attribute_position),
+                sort_position: new.ranking_rules.sort_position.or(self.ranking_rules.sort_position),
+                exactness_position: new
+                    .ranking_rules
+                    .exactness_position
+                    .or(self.ranking_rules.exactness_position),
+                values: new.ranking_rules.values.or(self.ranking_rules.values),
+            },
+            searchable_attributes: SearchableAttributesAnalytics {
+                total: new.searchable_attributes.total.or(self.searchable_attributes.total),
+                with_wildcard: new
+                    .searchable_attributes
+                    .with_wildcard
+                    .or(self.searchable_attributes.with_wildcard),
+            },
+            displayed_attributes: DisplayedAttributesAnalytics {
+                total: new.displayed_attributes.total.or(self.displayed_attributes.total),
+                with_wildcard: new
+                    .displayed_attributes
+                    .with_wildcard
+                    .or(self.displayed_attributes.with_wildcard),
+            },
+            sortable_attributes: SortableAttributesAnalytics {
+                total: new.sortable_attributes.total.or(self.sortable_attributes.total),
+                has_geo: new.sortable_attributes.has_geo.or(self.sortable_attributes.has_geo),
+            },
+            filterable_attributes: FilterableAttributesAnalytics {
+                total: new.filterable_attributes.total.or(self.filterable_attributes.total),
+                has_geo: new.filterable_attributes.has_geo.or(self.filterable_attributes.has_geo),
+            },
+            distinct_attribute: DistinctAttributeAnalytics {
+                set: self.distinct_attribute.set | new.distinct_attribute.set,
+            },
+            proximity_precision: ProximityPrecisionAnalytics {
+                set: self.proximity_precision.set | new.proximity_precision.set,
+                value: new.proximity_precision.value.or(self.proximity_precision.value),
+            },
+            typo_tolerance: TypoToleranceAnalytics {
+                enabled: new.typo_tolerance.enabled.or(self.typo_tolerance.enabled),
+                disable_on_attributes: new
+                    .typo_tolerance
+                    .disable_on_attributes
+                    .or(self.typo_tolerance.disable_on_attributes),
+                disable_on_words: new
+                    .typo_tolerance
+                    .disable_on_words
+                    .or(self.typo_tolerance.disable_on_words),
+                min_word_size_for_one_typo: new
+                    .typo_tolerance
+                    .min_word_size_for_one_typo
+                    .or(self.typo_tolerance.min_word_size_for_one_typo),
+                min_word_size_for_two_typos: new
+                    .typo_tolerance
+                    .min_word_size_for_two_typos
+                    .or(self.typo_tolerance.min_word_size_for_two_typos),
+            },
+            faceting: FacetingAnalytics {
+                max_values_per_facet: new
+                    .faceting
+                    .max_values_per_facet
+                    .or(self.faceting.max_values_per_facet),
+                sort_facet_values_by_star_count: new
+                    .faceting
+                    .sort_facet_values_by_star_count
+                    .or(self.faceting.sort_facet_values_by_star_count),
+                sort_facet_values_by_total: new
+                    .faceting
+                    .sort_facet_values_by_total
+                    .or(self.faceting.sort_facet_values_by_total),
+            },
+            pagination: PaginationAnalytics {
+                max_total_hits: new.pagination.max_total_hits.or(self.pagination.max_total_hits),
+            },
+            stop_words: StopWordsAnalytics {
+                total: new.stop_words.total.or(self.stop_words.total),
+            },
+            synonyms: SynonymsAnalytics { total: new.synonyms.total.or(self.synonyms.total) },
+            embedders: EmbeddersAnalytics {
+                total: new.embedders.total.or(self.embedders.total),
+                sources: match (self.embedders.sources, new.embedders.sources) {
+                    (None, None) => None,
+                    (Some(sources), None) | (None, Some(sources)) => Some(sources),
+                    (Some(this), Some(other)) => Some(this.union(&other).cloned().collect()),
+                },
+                document_template_used: match (
+                    self.embedders.document_template_used,
+                    new.embedders.document_template_used,
+                ) {
+                    (None, None) => None,
+                    (Some(used), None) | (None, Some(used)) => Some(used),
+                    (Some(this), Some(other)) => Some(this | other),
+                },
+                document_template_max_bytes: match (
+                    self.embedders.document_template_max_bytes,
+                    new.embedders.document_template_max_bytes,
+                ) {
+                    (None, None) => None,
+                    (Some(bytes), None) | (None, Some(bytes)) => Some(bytes),
+                    (Some(this), Some(other)) => Some(this.max(other)),
+                },
+                binary_quantization_used: match (
+                    self.embedders.binary_quantization_used,
+                    new.embedders.binary_quantization_used,
+                ) {
+                    (None, None) => None,
+                    (Some(bq), None) | (None, Some(bq)) => Some(bq),
+                    (Some(this), Some(other)) => Some(this | other),
+                },
+            },
+            search_cutoff_ms: SearchCutoffMsAnalytics {
+                search_cutoff_ms: new
+                    .search_cutoff_ms
+                    .search_cutoff_ms
+                    .or(self.search_cutoff_ms.search_cutoff_ms),
+            },
+            locales: LocalesAnalytics { locales: new.locales.locales.or(self.locales.locales) },
+            dictionary: DictionaryAnalytics {
+                total: new.dictionary.total.or(self.dictionary.total),
+            },
+            separator_tokens: SeparatorTokensAnalytics {
+                total: new.non_separator_tokens.total.or(self.separator_tokens.total),
+            },
+            non_separator_tokens: NonSeparatorTokensAnalytics {
+                total: new.non_separator_tokens.total.or(self.non_separator_tokens.total),
+            },
+        })
+    }
+
+    fn into_event(self: Box<Self>) -> serde_json::Value {
+        serde_json::to_value(*self).unwrap_or_default()
+    }
+}
+
+#[derive(Serialize, Default)]
+pub struct RankingRulesAnalytics {
+    pub words_position: Option<usize>,
+    pub typo_position: Option<usize>,
+    pub proximity_position: Option<usize>,
+    pub attribute_position: Option<usize>,
+    pub sort_position: Option<usize>,
+    pub exactness_position: Option<usize>,
+    pub values: Option<String>,
+}
+
+impl RankingRulesAnalytics {
+    pub fn new(rr: Option<&Vec<RankingRuleView>>) -> Self {
+        RankingRulesAnalytics {
+            words_position: rr.as_ref().and_then(|rr| {
+                rr.iter()
+                    .position(|s| matches!(s, meilisearch_types::settings::RankingRuleView::Words))
+            }),
+            typo_position: rr.as_ref().and_then(|rr| {
+                rr.iter()
+                    .position(|s| matches!(s, meilisearch_types::settings::RankingRuleView::Typo))
+            }),
+            proximity_position: rr.as_ref().and_then(|rr| {
+                rr.iter().position(|s| {
+                    matches!(s, meilisearch_types::settings::RankingRuleView::Proximity)
+                })
+            }),
+            attribute_position: rr.as_ref().and_then(|rr| {
+                rr.iter().position(|s| {
+                    matches!(s, meilisearch_types::settings::RankingRuleView::Attribute)
+                })
+            }),
+            sort_position: rr.as_ref().and_then(|rr| {
+                rr.iter()
+                    .position(|s| matches!(s, meilisearch_types::settings::RankingRuleView::Sort))
+            }),
+            exactness_position: rr.as_ref().and_then(|rr| {
+                rr.iter().position(|s| {
+                    matches!(s, meilisearch_types::settings::RankingRuleView::Exactness)
+                })
+            }),
+            values: rr.as_ref().map(|rr| {
+                rr.iter()
+                    .filter(|s| {
+                        matches!(
+                            s,
+                            meilisearch_types::settings::RankingRuleView::Asc(_)
+                                | meilisearch_types::settings::RankingRuleView::Desc(_)
+                        )
+                    })
+                    .map(|x| x.to_string())
+                    .collect::<Vec<_>>()
+                    .join(", ")
+            }),
+        }
+    }
+
+    pub fn into_settings(self) -> SettingsAnalytics {
+        SettingsAnalytics { ranking_rules: self, ..Default::default() }
+    }
+}
+
+#[derive(Serialize, Default)]
+pub struct SearchableAttributesAnalytics {
+    pub total: Option<usize>,
+    pub with_wildcard: Option<bool>,
+}
+
+impl SearchableAttributesAnalytics {
+    pub fn new(setting: Option<&Vec<String>>) -> Self {
+        Self {
+            total: setting.as_ref().map(|searchable| searchable.len()),
+            with_wildcard: setting
+                .as_ref()
+                .map(|searchable| searchable.iter().any(|searchable| searchable == "*")),
+        }
+    }
+
+    pub fn into_settings(self) -> SettingsAnalytics {
+        SettingsAnalytics { searchable_attributes: self, ..Default::default() }
+    }
+}
+
+#[derive(Serialize, Default)]
+pub struct DisplayedAttributesAnalytics {
+    pub total: Option<usize>,
+    pub with_wildcard: Option<bool>,
+}
+
+impl DisplayedAttributesAnalytics {
+    pub fn new(displayed: Option<&Vec<String>>) -> Self {
+        Self {
+            total: displayed.as_ref().map(|displayed| displayed.len()),
+            with_wildcard: displayed
+                .as_ref()
+                .map(|displayed| displayed.iter().any(|displayed| displayed == "*")),
+        }
+    }
+
+    pub fn into_settings(self) -> SettingsAnalytics {
+        SettingsAnalytics { displayed_attributes: self, ..Default::default() }
+    }
+}
+
+#[derive(Serialize, Default)]
+pub struct SortableAttributesAnalytics {
+    pub total: Option<usize>,
+    pub has_geo: Option<bool>,
+}
+
+impl SortableAttributesAnalytics {
+    pub fn new(setting: Option<&BTreeSet<String>>) -> Self {
+        Self {
+            total: setting.as_ref().map(|sort| sort.len()),
+            has_geo: setting.as_ref().map(|sort| sort.contains("_geo")),
+        }
+    }
+
+    pub fn into_settings(self) -> SettingsAnalytics {
+        SettingsAnalytics { sortable_attributes: self, ..Default::default() }
+    }
+}
+
+#[derive(Serialize, Default)]
+pub struct FilterableAttributesAnalytics {
+    pub total: Option<usize>,
+    pub has_geo: Option<bool>,
+}
+
+impl FilterableAttributesAnalytics {
+    pub fn new(setting: Option<&BTreeSet<String>>) -> Self {
+        Self {
+            total: setting.as_ref().map(|filter| filter.len()),
+            has_geo: setting.as_ref().map(|filter| filter.contains("_geo")),
+        }
+    }
+
+    pub fn into_settings(self) -> SettingsAnalytics {
+        SettingsAnalytics { filterable_attributes: self, ..Default::default() }
+    }
+}
+
+#[derive(Serialize, Default)]
+pub struct DistinctAttributeAnalytics {
+    pub set: bool,
+}
+
+impl DistinctAttributeAnalytics {
+    pub fn new(distinct: Option<&String>) -> Self {
+        Self { set: distinct.is_some() }
+    }
+
+    pub fn into_settings(self) -> SettingsAnalytics {
+        SettingsAnalytics { distinct_attribute: self, ..Default::default() }
+    }
+}
+
+#[derive(Serialize, Default)]
+pub struct ProximityPrecisionAnalytics {
+    pub set: bool,
+    pub value: Option<ProximityPrecisionView>,
+}
+
+impl ProximityPrecisionAnalytics {
+    pub fn new(precision: Option<&ProximityPrecisionView>) -> Self {
+        Self { set: precision.is_some(), value: precision.cloned() }
+    }
+
+    pub fn into_settings(self) -> SettingsAnalytics {
+        SettingsAnalytics { proximity_precision: self, ..Default::default() }
+    }
+}
+
+#[derive(Serialize, Default)]
+pub struct TypoToleranceAnalytics {
+    pub enabled: Option<bool>,
+    pub disable_on_attributes: Option<bool>,
+    pub disable_on_words: Option<bool>,
+    pub min_word_size_for_one_typo: Option<u8>,
+    pub min_word_size_for_two_typos: Option<u8>,
+}
+
+impl TypoToleranceAnalytics {
+    pub fn new(setting: Option<&TypoSettings>) -> Self {
+        Self {
+            enabled: setting.as_ref().map(|s| !matches!(s.enabled, Setting::Set(false))),
+            disable_on_attributes: setting
+                .as_ref()
+                .and_then(|s| s.disable_on_attributes.as_ref().set().map(|m| !m.is_empty())),
+            disable_on_words: setting
+                .as_ref()
+                .and_then(|s| s.disable_on_words.as_ref().set().map(|m| !m.is_empty())),
+            min_word_size_for_one_typo: setting
+                .as_ref()
+                .and_then(|s| s.min_word_size_for_typos.as_ref().set().map(|s| s.one_typo.set()))
+                .flatten(),
+            min_word_size_for_two_typos: setting
+                .as_ref()
+                .and_then(|s| s.min_word_size_for_typos.as_ref().set().map(|s| s.two_typos.set()))
+                .flatten(),
+        }
+    }
+    pub fn into_settings(self) -> SettingsAnalytics {
+        SettingsAnalytics { typo_tolerance: self, ..Default::default() }
+    }
+}
+
+#[derive(Serialize, Default)]
+pub struct FacetingAnalytics {
+    pub max_values_per_facet: Option<usize>,
+    pub sort_facet_values_by_star_count: Option<bool>,
+    pub sort_facet_values_by_total: Option<usize>,
+}
+
+impl FacetingAnalytics {
+    pub fn new(setting: Option<&FacetingSettings>) -> Self {
+        Self {
+            max_values_per_facet: setting.as_ref().and_then(|s| s.max_values_per_facet.set()),
+            sort_facet_values_by_star_count: setting.as_ref().and_then(|s| {
+                s.sort_facet_values_by
+                    .as_ref()
+                    .set()
+                    .map(|s| s.iter().any(|(k, v)| k == "*" && v == &FacetValuesSort::Count))
+            }),
+            sort_facet_values_by_total: setting
+                .as_ref()
+                .and_then(|s| s.sort_facet_values_by.as_ref().set().map(|s| s.len())),
+        }
+    }
+
+    pub fn into_settings(self) -> SettingsAnalytics {
+        SettingsAnalytics { faceting: self, ..Default::default() }
+    }
+}
+
+#[derive(Serialize, Default)]
+pub struct PaginationAnalytics {
+    pub max_total_hits: Option<usize>,
+}
+
+impl PaginationAnalytics {
+    pub fn new(setting: Option<&PaginationSettings>) -> Self {
+        Self { max_total_hits: setting.as_ref().and_then(|s| s.max_total_hits.set()) }
+    }
+
+    pub fn into_settings(self) -> SettingsAnalytics {
+        SettingsAnalytics { pagination: self, ..Default::default() }
+    }
+}
+
+#[derive(Serialize, Default)]
+pub struct StopWordsAnalytics {
+    pub total: Option<usize>,
+}
+
+impl StopWordsAnalytics {
+    pub fn new(stop_words: Option<&BTreeSet<String>>) -> Self {
+        Self { total: stop_words.as_ref().map(|stop_words| stop_words.len()) }
+    }
+
+    pub fn into_settings(self) -> SettingsAnalytics {
+        SettingsAnalytics { stop_words: self, ..Default::default() }
+    }
+}
+
+#[derive(Serialize, Default)]
+pub struct SynonymsAnalytics {
+    pub total: Option<usize>,
+}
+
+impl SynonymsAnalytics {
+    pub fn new(synonyms: Option<&BTreeMap<String, Vec<String>>>) -> Self {
+        Self { total: synonyms.as_ref().map(|synonyms| synonyms.len()) }
+    }
+
+    pub fn into_settings(self) -> SettingsAnalytics {
+        SettingsAnalytics { synonyms: self, ..Default::default() }
+    }
+}
+
+#[derive(Serialize, Default)]
+pub struct EmbeddersAnalytics {
+    // last
+    pub total: Option<usize>,
+    // Merge the sources
+    pub sources: Option<HashSet<String>>,
+    // |=
+    pub document_template_used: Option<bool>,
+    // max
+    pub document_template_max_bytes: Option<usize>,
+    // |=
+    pub binary_quantization_used: Option<bool>,
+}
+
+impl EmbeddersAnalytics {
+    pub fn new(setting: Option<&BTreeMap<String, Setting<EmbeddingSettings>>>) -> Self {
+        let mut sources = std::collections::HashSet::new();
+
+        if let Some(s) = &setting {
+            for source in s
+                .values()
+                .filter_map(|config| config.clone().set())
+                .filter_map(|config| config.source.set())
+            {
+                use meilisearch_types::milli::vector::settings::EmbedderSource;
+                match source {
+                    EmbedderSource::OpenAi => sources.insert("openAi".to_string()),
+                    EmbedderSource::HuggingFace => sources.insert("huggingFace".to_string()),
+                    EmbedderSource::UserProvided => sources.insert("userProvided".to_string()),
+                    EmbedderSource::Ollama => sources.insert("ollama".to_string()),
+                    EmbedderSource::Rest => sources.insert("rest".to_string()),
+                };
+            }
+        };
+
+        Self {
+            total: setting.as_ref().map(|s| s.len()),
+            sources: Some(sources),
+            document_template_used: setting.as_ref().map(|map| {
+                map.values()
+                    .filter_map(|config| config.clone().set())
+                    .any(|config| config.document_template.set().is_some())
+            }),
+            document_template_max_bytes: setting.as_ref().and_then(|map| {
+                map.values()
+                    .filter_map(|config| config.clone().set())
+                    .filter_map(|config| config.document_template_max_bytes.set())
+                    .max()
+            }),
+            binary_quantization_used: setting.as_ref().map(|map| {
+                map.values()
+                    .filter_map(|config| config.clone().set())
+                    .any(|config| config.binary_quantized.set().is_some())
+            }),
+        }
+    }
+
+    pub fn into_settings(self) -> SettingsAnalytics {
+        SettingsAnalytics { embedders: self, ..Default::default() }
+    }
+}
+
+#[derive(Serialize, Default)]
+#[serde(transparent)]
+pub struct SearchCutoffMsAnalytics {
+    pub search_cutoff_ms: Option<u64>,
+}
+
+impl SearchCutoffMsAnalytics {
+    pub fn new(setting: Option<&u64>) -> Self {
+        Self { search_cutoff_ms: setting.copied() }
+    }
+
+    pub fn into_settings(self) -> SettingsAnalytics {
+        SettingsAnalytics { search_cutoff_ms: self, ..Default::default() }
+    }
+}
+
+#[derive(Serialize, Default)]
+#[serde(transparent)]
+pub struct LocalesAnalytics {
+    pub locales: Option<BTreeSet<Locale>>,
+}
+
+impl LocalesAnalytics {
+    pub fn new(rules: Option<&Vec<LocalizedAttributesRuleView>>) -> Self {
+        LocalesAnalytics {
+            locales: rules.as_ref().map(|rules| {
+                rules
+                    .iter()
+                    .flat_map(|rule| rule.locales.iter().cloned())
+                    .collect::<std::collections::BTreeSet<_>>()
+            }),
+        }
+    }
+
+    pub fn into_settings(self) -> SettingsAnalytics {
+        SettingsAnalytics { locales: self, ..Default::default() }
+    }
+}
+
+#[derive(Serialize, Default)]
+pub struct DictionaryAnalytics {
+    pub total: Option<usize>,
+}
+
+impl DictionaryAnalytics {
+    pub fn new(dictionary: Option<&BTreeSet<String>>) -> Self {
+        Self { total: dictionary.as_ref().map(|dictionary| dictionary.len()) }
+    }
+
+    pub fn into_settings(self) -> SettingsAnalytics {
+        SettingsAnalytics { dictionary: self, ..Default::default() }
+    }
+}
+
+#[derive(Serialize, Default)]
+pub struct SeparatorTokensAnalytics {
+    pub total: Option<usize>,
+}
+
+impl SeparatorTokensAnalytics {
+    pub fn new(separator_tokens: Option<&BTreeSet<String>>) -> Self {
+        Self { total: separator_tokens.as_ref().map(|separator_tokens| separator_tokens.len()) }
+    }
+
+    pub fn into_settings(self) -> SettingsAnalytics {
+        SettingsAnalytics { separator_tokens: self, ..Default::default() }
+    }
+}
+
+#[derive(Serialize, Default)]
+pub struct NonSeparatorTokensAnalytics {
+    pub total: Option<usize>,
+}
+
+impl NonSeparatorTokensAnalytics {
+    pub fn new(non_separator_tokens: Option<&BTreeSet<String>>) -> Self {
+        Self {
+            total: non_separator_tokens
+                .as_ref()
+                .map(|non_separator_tokens| non_separator_tokens.len()),
+        }
+    }
+
+    pub fn into_settings(self) -> SettingsAnalytics {
+        SettingsAnalytics { non_separator_tokens: self, ..Default::default() }
+    }
+}

crates/meilisearch/src/routes/indexes/similar.rs

@@ -0,0 +1,201 @@
+use actix_web::web::{self, Data};
+use actix_web::{HttpRequest, HttpResponse};
+use deserr::actix_web::{AwebJson, AwebQueryParameter};
+use index_scheduler::IndexScheduler;
+use meilisearch_types::deserr::query_params::Param;
+use meilisearch_types::deserr::{DeserrJsonError, DeserrQueryParamError};
+use meilisearch_types::error::deserr_codes::*;
+use meilisearch_types::error::{ErrorCode as _, ResponseError};
+use meilisearch_types::index_uid::IndexUid;
+use meilisearch_types::keys::actions;
+use meilisearch_types::serde_cs::vec::CS;
+use serde_json::Value;
+use tracing::debug;
+
+use super::ActionPolicy;
+use crate::analytics::Analytics;
+use crate::extractors::authentication::GuardedData;
+use crate::extractors::sequential_extractor::SeqHandler;
+use crate::routes::indexes::similar_analytics::{SimilarAggregator, SimilarGET, SimilarPOST};
+use crate::search::{
+    add_search_rules, perform_similar, RankingScoreThresholdSimilar, RetrieveVectors, SearchKind,
+    SimilarQuery, SimilarResult, DEFAULT_SEARCH_LIMIT, DEFAULT_SEARCH_OFFSET,
+};
+
+pub fn configure(cfg: &mut web::ServiceConfig) {
+    cfg.service(
+        web::resource("")
+            .route(web::get().to(SeqHandler(similar_get)))
+            .route(web::post().to(SeqHandler(similar_post))),
+    );
+}
+
+pub async fn similar_get(
+    index_scheduler: GuardedData<ActionPolicy<{ actions::SEARCH }>, Data<IndexScheduler>>,
+    index_uid: web::Path<String>,
+    params: AwebQueryParameter<SimilarQueryGet, DeserrQueryParamError>,
+    req: HttpRequest,
+    analytics: web::Data<Analytics>,
+) -> Result<HttpResponse, ResponseError> {
+    let index_uid = IndexUid::try_from(index_uid.into_inner())?;
+
+    let query = params.0.try_into()?;
+
+    let mut aggregate = SimilarAggregator::<SimilarGET>::from_query(&query);
+
+    debug!(parameters = ?query, "Similar get");
+
+    let similar = similar(index_scheduler, index_uid, query).await;
+
+    if let Ok(similar) = &similar {
+        aggregate.succeed(similar);
+    }
+    analytics.publish(aggregate, &req);
+
+    let similar = similar?;
+
+    debug!(returns = ?similar, "Similar get");
+    Ok(HttpResponse::Ok().json(similar))
+}
+
+pub async fn similar_post(
+    index_scheduler: GuardedData<ActionPolicy<{ actions::SEARCH }>, Data<IndexScheduler>>,
+    index_uid: web::Path<String>,
+    params: AwebJson<SimilarQuery, DeserrJsonError>,
+    req: HttpRequest,
+    analytics: web::Data<Analytics>,
+) -> Result<HttpResponse, ResponseError> {
+    let index_uid = IndexUid::try_from(index_uid.into_inner())?;
+
+    let query = params.into_inner();
+    debug!(parameters = ?query, "Similar post");
+
+    let mut aggregate = SimilarAggregator::<SimilarPOST>::from_query(&query);
+
+    let similar = similar(index_scheduler, index_uid, query).await;
+
+    if let Ok(similar) = &similar {
+        aggregate.succeed(similar);
+    }
+    analytics.publish(aggregate, &req);
+
+    let similar = similar?;
+
+    debug!(returns = ?similar, "Similar post");
+    Ok(HttpResponse::Ok().json(similar))
+}
+
+async fn similar(
+    index_scheduler: GuardedData<ActionPolicy<{ actions::SEARCH }>, Data<IndexScheduler>>,
+    index_uid: IndexUid,
+    mut query: SimilarQuery,
+) -> Result<SimilarResult, ResponseError> {
+    let features = index_scheduler.features();
+
+    features.check_vector("Using the similar API")?;
+
+    let retrieve_vectors = RetrieveVectors::new(query.retrieve_vectors, features)?;
+
+    // Tenant token search_rules.
+    if let Some(search_rules) = index_scheduler.filters().get_index_search_rules(&index_uid) {
+        add_search_rules(&mut query.filter, search_rules);
+    }
+
+    let index = index_scheduler.index(&index_uid)?;
+
+    let (embedder_name, embedder, quantized) =
+        SearchKind::embedder(&index_scheduler, &index, &query.embedder, None)?;
+
+    tokio::task::spawn_blocking(move || {
+        perform_similar(
+            &index,
+            query,
+            embedder_name,
+            embedder,
+            quantized,
+            retrieve_vectors,
+            index_scheduler.features(),
+        )
+    })
+    .await?
+}
+
+#[derive(Debug, deserr::Deserr)]
+#[deserr(error = DeserrQueryParamError, rename_all = camelCase, deny_unknown_fields)]
+pub struct SimilarQueryGet {
+    #[deserr(error = DeserrQueryParamError<InvalidSimilarId>)]
+    id: Param<String>,
+    #[deserr(default = Param(DEFAULT_SEARCH_OFFSET()), error = DeserrQueryParamError<InvalidSimilarOffset>)]
+    offset: Param<usize>,
+    #[deserr(default = Param(DEFAULT_SEARCH_LIMIT()), error = DeserrQueryParamError<InvalidSimilarLimit>)]
+    limit: Param<usize>,
+    #[deserr(default, error = DeserrQueryParamError<InvalidSimilarAttributesToRetrieve>)]
+    attributes_to_retrieve: Option<CS<String>>,
+    #[deserr(default, error = DeserrQueryParamError<InvalidSimilarRetrieveVectors>)]
+    retrieve_vectors: Param<bool>,
+    #[deserr(default, error = DeserrQueryParamError<InvalidSimilarFilter>)]
+    filter: Option<String>,
+    #[deserr(default, error = DeserrQueryParamError<InvalidSimilarShowRankingScore>)]
+    show_ranking_score: Param<bool>,
+    #[deserr(default, error = DeserrQueryParamError<InvalidSimilarShowRankingScoreDetails>)]
+    show_ranking_score_details: Param<bool>,
+    #[deserr(default, error = DeserrQueryParamError<InvalidSimilarRankingScoreThreshold>, default)]
+    pub ranking_score_threshold: Option<RankingScoreThresholdGet>,
+    #[deserr(error = DeserrQueryParamError<InvalidEmbedder>)]
+    pub embedder: String,
+}
+
+#[derive(Debug, Clone, Copy, PartialEq, deserr::Deserr)]
+#[deserr(try_from(String) = TryFrom::try_from -> InvalidSimilarRankingScoreThreshold)]
+pub struct RankingScoreThresholdGet(RankingScoreThresholdSimilar);
+
+impl std::convert::TryFrom<String> for RankingScoreThresholdGet {
+    type Error = InvalidSimilarRankingScoreThreshold;
+
+    fn try_from(s: String) -> Result<Self, Self::Error> {
+        let f: f64 = s.parse().map_err(|_| InvalidSimilarRankingScoreThreshold)?;
+        Ok(RankingScoreThresholdGet(RankingScoreThresholdSimilar::try_from(f)?))
+    }
+}
+
+impl TryFrom<SimilarQueryGet> for SimilarQuery {
+    type Error = ResponseError;
+
+    fn try_from(
+        SimilarQueryGet {
+            id,
+            offset,
+            limit,
+            attributes_to_retrieve,
+            retrieve_vectors,
+            filter,
+            show_ranking_score,
+            show_ranking_score_details,
+            embedder,
+            ranking_score_threshold,
+        }: SimilarQueryGet,
+    ) -> Result<Self, Self::Error> {
+        let filter = match filter {
+            Some(f) => match serde_json::from_str(&f) {
+                Ok(v) => Some(v),
+                _ => Some(Value::String(f)),
+            },
+            None => None,
+        };
+
+        Ok(SimilarQuery {
+            id: id.0.try_into().map_err(|code: InvalidSimilarId| {
+                ResponseError::from_msg(code.to_string(), code.error_code())
+            })?,
+            offset: offset.0,
+            limit: limit.0,
+            filter,
+            embedder,
+            attributes_to_retrieve: attributes_to_retrieve.map(|o| o.into_iter().collect()),
+            retrieve_vectors: retrieve_vectors.0,
+            show_ranking_score: show_ranking_score.0,
+            show_ranking_score_details: show_ranking_score_details.0,
+            ranking_score_threshold: ranking_score_threshold.map(|x| x.0),
+        })
+    }
+}

crates/meilisearch/src/routes/indexes/similar_analytics.rs

@@ -0,0 +1,235 @@
+use std::collections::{BinaryHeap, HashMap};
+
+use once_cell::sync::Lazy;
+use regex::Regex;
+use serde_json::{json, Value};
+
+use crate::{
+    aggregate_methods,
+    analytics::{Aggregate, AggregateMethod},
+    search::{SimilarQuery, SimilarResult},
+};
+
+aggregate_methods!(
+    SimilarPOST => "Similar POST",
+    SimilarGET => "Similar GET",
+);
+
+#[derive(Default)]
+pub struct SimilarAggregator<Method: AggregateMethod> {
+    // requests
+    total_received: usize,
+    total_succeeded: usize,
+    time_spent: BinaryHeap<usize>,
+
+    // filter
+    filter_with_geo_radius: bool,
+    filter_with_geo_bounding_box: bool,
+    // every time a request has a filter, this field must be incremented by the number of terms it contains
+    filter_sum_of_criteria_terms: usize,
+    // every time a request has a filter, this field must be incremented by one
+    filter_total_number_of_criteria: usize,
+    used_syntax: HashMap<String, usize>,
+
+    // Whether a non-default embedder was specified
+    retrieve_vectors: bool,
+
+    // pagination
+    max_limit: usize,
+    max_offset: usize,
+
+    // formatting
+    max_attributes_to_retrieve: usize,
+
+    // scoring
+    show_ranking_score: bool,
+    show_ranking_score_details: bool,
+    ranking_score_threshold: bool,
+
+    marker: std::marker::PhantomData<Method>,
+}
+
+impl<Method: AggregateMethod> SimilarAggregator<Method> {
+    #[allow(clippy::field_reassign_with_default)]
+    pub fn from_query(query: &SimilarQuery) -> Self {
+        let SimilarQuery {
+            id: _,
+            embedder: _,
+            offset,
+            limit,
+            attributes_to_retrieve: _,
+            retrieve_vectors,
+            show_ranking_score,
+            show_ranking_score_details,
+            filter,
+            ranking_score_threshold,
+        } = query;
+
+        let mut ret = Self::default();
+
+        ret.total_received = 1;
+
+        if let Some(ref filter) = filter {
+            static RE: Lazy<Regex> = Lazy::new(|| Regex::new("AND | OR").unwrap());
+            ret.filter_total_number_of_criteria = 1;
+
+            let syntax = match filter {
+                Value::String(_) => "string".to_string(),
+                Value::Array(values) => {
+                    if values.iter().map(|v| v.to_string()).any(|s| RE.is_match(&s)) {
+                        "mixed".to_string()
+                    } else {
+                        "array".to_string()
+                    }
+                }
+                _ => "none".to_string(),
+            };
+            // convert the string to a HashMap
+            ret.used_syntax.insert(syntax, 1);
+
+            let stringified_filters = filter.to_string();
+            ret.filter_with_geo_radius = stringified_filters.contains("_geoRadius(");
+            ret.filter_with_geo_bounding_box = stringified_filters.contains("_geoBoundingBox(");
+            ret.filter_sum_of_criteria_terms = RE.split(&stringified_filters).count();
+        }
+
+        ret.max_limit = *limit;
+        ret.max_offset = *offset;
+
+        ret.show_ranking_score = *show_ranking_score;
+        ret.show_ranking_score_details = *show_ranking_score_details;
+        ret.ranking_score_threshold = ranking_score_threshold.is_some();
+
+        ret.retrieve_vectors = *retrieve_vectors;
+
+        ret
+    }
+
+    pub fn succeed(&mut self, result: &SimilarResult) {
+        let SimilarResult { id: _, hits: _, processing_time_ms, hits_info: _ } = result;
+
+        self.total_succeeded = self.total_succeeded.saturating_add(1);
+
+        self.time_spent.push(*processing_time_ms as usize);
+    }
+}
+
+impl<Method: AggregateMethod> Aggregate for SimilarAggregator<Method> {
+    fn event_name(&self) -> &'static str {
+        Method::event_name()
+    }
+
+    /// Aggregate one [SimilarAggregator] into another.
+    fn aggregate(mut self: Box<Self>, new: Box<Self>) -> Box<Self> {
+        let Self {
+            total_received,
+            total_succeeded,
+            mut time_spent,
+            filter_with_geo_radius,
+            filter_with_geo_bounding_box,
+            filter_sum_of_criteria_terms,
+            filter_total_number_of_criteria,
+            used_syntax,
+            max_limit,
+            max_offset,
+            max_attributes_to_retrieve,
+            show_ranking_score,
+            show_ranking_score_details,
+            ranking_score_threshold,
+            retrieve_vectors,
+            marker: _,
+        } = *new;
+
+        // request
+        self.total_received = self.total_received.saturating_add(total_received);
+        self.total_succeeded = self.total_succeeded.saturating_add(total_succeeded);
+        self.time_spent.append(&mut time_spent);
+
+        // filter
+        self.filter_with_geo_radius |= filter_with_geo_radius;
+        self.filter_with_geo_bounding_box |= filter_with_geo_bounding_box;
+        self.filter_sum_of_criteria_terms =
+            self.filter_sum_of_criteria_terms.saturating_add(filter_sum_of_criteria_terms);
+        self.filter_total_number_of_criteria =
+            self.filter_total_number_of_criteria.saturating_add(filter_total_number_of_criteria);
+        for (key, value) in used_syntax.into_iter() {
+            let used_syntax = self.used_syntax.entry(key).or_insert(0);
+            *used_syntax = used_syntax.saturating_add(value);
+        }
+
+        self.retrieve_vectors |= retrieve_vectors;
+
+        // pagination
+        self.max_limit = self.max_limit.max(max_limit);
+        self.max_offset = self.max_offset.max(max_offset);
+
+        // formatting
+        self.max_attributes_to_retrieve =
+            self.max_attributes_to_retrieve.max(max_attributes_to_retrieve);
+
+        // scoring
+        self.show_ranking_score |= show_ranking_score;
+        self.show_ranking_score_details |= show_ranking_score_details;
+        self.ranking_score_threshold |= ranking_score_threshold;
+
+        self
+    }
+
+    fn into_event(self: Box<Self>) -> serde_json::Value {
+        let Self {
+            total_received,
+            total_succeeded,
+            time_spent,
+            filter_with_geo_radius,
+            filter_with_geo_bounding_box,
+            filter_sum_of_criteria_terms,
+            filter_total_number_of_criteria,
+            used_syntax,
+            max_limit,
+            max_offset,
+            max_attributes_to_retrieve,
+            show_ranking_score,
+            show_ranking_score_details,
+            ranking_score_threshold,
+            retrieve_vectors,
+            marker: _,
+        } = *self;
+
+        // we get all the values in a sorted manner
+        let time_spent = time_spent.into_sorted_vec();
+        // the index of the 99th percentage of value
+        let percentile_99th = time_spent.len() * 99 / 100;
+        // We are only interested by the slowest value of the 99th fastest results
+        let time_spent = time_spent.get(percentile_99th);
+
+        json!({
+            "requests": {
+                "99th_response_time": time_spent.map(|t| format!("{:.2}", t)),
+                "total_succeeded": total_succeeded,
+                "total_failed": total_received.saturating_sub(total_succeeded), // just to be sure we never panics
+                "total_received": total_received,
+            },
+            "filter": {
+               "with_geoRadius": filter_with_geo_radius,
+               "with_geoBoundingBox": filter_with_geo_bounding_box,
+               "avg_criteria_number": format!("{:.2}", filter_sum_of_criteria_terms as f64 / filter_total_number_of_criteria as f64),
+               "most_used_syntax": used_syntax.iter().max_by_key(|(_, v)| *v).map(|(k, _)| json!(k)).unwrap_or_else(|| json!(null)),
+            },
+            "vector": {
+                "retrieve_vectors": retrieve_vectors,
+            },
+            "pagination": {
+               "max_limit": max_limit,
+               "max_offset": max_offset,
+            },
+            "formatting": {
+                "max_attributes_to_retrieve": max_attributes_to_retrieve,
+            },
+            "scoring": {
+                "show_ranking_score": show_ranking_score,
+                "show_ranking_score_details": show_ranking_score_details,
+                "ranking_score_threshold": ranking_score_threshold,
+            }
+        })
+    }
+}

crates/meilisearch/src/routes/logs.rs

@@ -0,0 +1,318 @@
+use std::convert::Infallible;
+use std::io::Write;
+use std::ops::ControlFlow;
+use std::pin::Pin;
+use std::str::FromStr;
+use std::sync::Arc;
+
+use actix_web::web::{Bytes, Data};
+use actix_web::{web, HttpResponse};
+use deserr::actix_web::AwebJson;
+use deserr::{DeserializeError, Deserr, ErrorKind, MergeWithError, ValuePointerRef};
+use futures_util::Stream;
+use index_scheduler::IndexScheduler;
+use meilisearch_types::deserr::DeserrJsonError;
+use meilisearch_types::error::deserr_codes::*;
+use meilisearch_types::error::{Code, ResponseError};
+use tokio::sync::mpsc;
+use tracing_subscriber::filter::Targets;
+use tracing_subscriber::Layer;
+
+use crate::error::MeilisearchHttpError;
+use crate::extractors::authentication::policies::*;
+use crate::extractors::authentication::GuardedData;
+use crate::extractors::sequential_extractor::SeqHandler;
+use crate::{LogRouteHandle, LogStderrHandle};
+
+pub fn configure(cfg: &mut web::ServiceConfig) {
+    cfg.service(
+        web::resource("stream")
+            .route(web::post().to(SeqHandler(get_logs)))
+            .route(web::delete().to(SeqHandler(cancel_logs))),
+    )
+    .service(web::resource("stderr").route(web::post().to(SeqHandler(update_stderr_target))));
+}
+
+#[derive(Debug, Default, Clone, Copy, Deserr, PartialEq, Eq)]
+#[deserr(rename_all = camelCase)]
+pub enum LogMode {
+    #[default]
+    Human,
+    Json,
+    Profile,
+}
+
+/// Simple wrapper around the `Targets` from `tracing_subscriber` to implement `MergeWithError` on it.
+#[derive(Clone, Debug)]
+struct MyTargets(Targets);
+
+/// Simple wrapper around the `ParseError` from `tracing_subscriber` to implement `MergeWithError` on it.
+#[derive(Debug, thiserror::Error)]
+enum MyParseError {
+    #[error(transparent)]
+    ParseError(#[from] tracing_subscriber::filter::ParseError),
+    #[error(
+        "Empty string is not a valid target. If you want to get no logs use `OFF`. Usage: `info`, `meilisearch=info`, or you can write multiple filters in one target: `index_scheduler=info,milli=trace`"
+    )]
+    Example,
+}
+
+impl FromStr for MyTargets {
+    type Err = MyParseError;
+
+    fn from_str(s: &str) -> Result<Self, Self::Err> {
+        if s.is_empty() {
+            Err(MyParseError::Example)
+        } else {
+            Ok(MyTargets(Targets::from_str(s).map_err(MyParseError::ParseError)?))
+        }
+    }
+}
+
+impl MergeWithError<MyParseError> for DeserrJsonError<BadRequest> {
+    fn merge(
+        _self_: Option<Self>,
+        other: MyParseError,
+        merge_location: ValuePointerRef,
+    ) -> ControlFlow<Self, Self> {
+        Self::error::<Infallible>(
+            None,
+            ErrorKind::Unexpected { msg: other.to_string() },
+            merge_location,
+        )
+    }
+}
+
+#[derive(Debug, Deserr)]
+#[deserr(error = DeserrJsonError, rename_all = camelCase, deny_unknown_fields, validate = validate_get_logs -> DeserrJsonError<InvalidSettingsTypoTolerance>)]
+pub struct GetLogs {
+    #[deserr(default = "info".parse().unwrap(), try_from(&String) = MyTargets::from_str -> DeserrJsonError<BadRequest>)]
+    target: MyTargets,
+
+    #[deserr(default, error = DeserrJsonError<BadRequest>)]
+    mode: LogMode,
+
+    #[deserr(default = false, error = DeserrJsonError<BadRequest>)]
+    profile_memory: bool,
+}
+
+fn validate_get_logs<E: DeserializeError>(
+    logs: GetLogs,
+    location: ValuePointerRef,
+) -> Result<GetLogs, E> {
+    if logs.profile_memory && logs.mode != LogMode::Profile {
+        Err(deserr::take_cf_content(E::error::<Infallible>(
+            None,
+            ErrorKind::Unexpected {
+                msg: format!("`profile_memory` can only be used while profiling code and is not compatible with the {:?} mode.", logs.mode),
+            },
+            location,
+        )))
+    } else {
+        Ok(logs)
+    }
+}
+
+struct LogWriter {
+    sender: mpsc::UnboundedSender<Vec<u8>>,
+}
+
+impl Write for LogWriter {
+    fn write(&mut self, buf: &[u8]) -> std::io::Result<usize> {
+        self.sender.send(buf.to_vec()).map_err(std::io::Error::other)?;
+        Ok(buf.len())
+    }
+
+    fn flush(&mut self) -> std::io::Result<()> {
+        Ok(())
+    }
+}
+
+struct HandleGuard {
+    /// We need to keep an handle on the logs to make it available again when the streamer is dropped
+    logs: Arc<LogRouteHandle>,
+}
+
+impl Drop for HandleGuard {
+    fn drop(&mut self) {
+        if let Err(e) = self.logs.modify(|layer| *layer.inner_mut() = None) {
+            tracing::error!("Could not free the logs route: {e}");
+        }
+    }
+}
+
+fn byte_stream(
+    receiver: mpsc::UnboundedReceiver<Vec<u8>>,
+    guard: HandleGuard,
+) -> impl futures_util::Stream<Item = Result<Bytes, ResponseError>> {
+    futures_util::stream::unfold((receiver, guard), move |(mut receiver, guard)| async move {
+        let vec = receiver.recv().await;
+
+        vec.map(From::from).map(Ok).map(|a| (a, (receiver, guard)))
+    })
+}
+
+type PinnedByteStream = Pin<Box<dyn Stream<Item = Result<Bytes, ResponseError>>>>;
+
+fn make_layer<
+    S: tracing::Subscriber + for<'span> tracing_subscriber::registry::LookupSpan<'span>,
+>(
+    opt: &GetLogs,
+    logs: Data<LogRouteHandle>,
+) -> (Box<dyn Layer<S> + Send + Sync>, PinnedByteStream) {
+    let guard = HandleGuard { logs: logs.into_inner() };
+    match opt.mode {
+        LogMode::Human => {
+            let (sender, receiver) = tokio::sync::mpsc::unbounded_channel();
+
+            let fmt_layer = tracing_subscriber::fmt::layer()
+                .with_writer(move || LogWriter { sender: sender.clone() })
+                .with_span_events(tracing_subscriber::fmt::format::FmtSpan::CLOSE);
+
+            let stream = byte_stream(receiver, guard);
+            (Box::new(fmt_layer) as Box<dyn Layer<S> + Send + Sync>, Box::pin(stream))
+        }
+        LogMode::Json => {
+            let (sender, receiver) = tokio::sync::mpsc::unbounded_channel();
+
+            let fmt_layer = tracing_subscriber::fmt::layer()
+                .with_writer(move || LogWriter { sender: sender.clone() })
+                .json()
+                .with_span_events(tracing_subscriber::fmt::format::FmtSpan::CLOSE);
+
+            let stream = byte_stream(receiver, guard);
+            (Box::new(fmt_layer) as Box<dyn Layer<S> + Send + Sync>, Box::pin(stream))
+        }
+        LogMode::Profile => {
+            let (trace, layer) = tracing_trace::Trace::new(opt.profile_memory);
+
+            let stream = entry_stream(trace, guard);
+
+            (Box::new(layer) as Box<dyn Layer<S> + Send + Sync>, Box::pin(stream))
+        }
+    }
+}
+
+fn entry_stream(
+    trace: tracing_trace::Trace,
+    guard: HandleGuard,
+) -> impl Stream<Item = Result<Bytes, ResponseError>> {
+    let receiver = trace.into_receiver();
+    let entry_buf = Vec::new();
+
+    futures_util::stream::unfold(
+        (receiver, entry_buf, guard),
+        move |(mut receiver, mut entry_buf, guard)| async move {
+            let mut bytes = Vec::new();
+
+            while bytes.len() < 8192 {
+                entry_buf.clear();
+
+                let Ok(count) = tokio::time::timeout(
+                    std::time::Duration::from_secs(1),
+                    receiver.recv_many(&mut entry_buf, 100),
+                )
+                .await
+                else {
+                    break;
+                };
+
+                if count == 0 {
+                    if !bytes.is_empty() {
+                        break;
+                    }
+
+                    // channel closed, exit
+                    return None;
+                }
+
+                for entry in &entry_buf {
+                    if let Err(error) = serde_json::to_writer(&mut bytes, entry) {
+                        tracing::error!(
+                            error = &error as &dyn std::error::Error,
+                            "deserializing entry"
+                        );
+                        return Some((
+                            Err(ResponseError::from_msg(
+                                format!("error deserializing entry: {error}"),
+                                Code::Internal,
+                            )),
+                            (receiver, entry_buf, guard),
+                        ));
+                    }
+                }
+            }
+
+            Some((Ok(bytes.into()), (receiver, entry_buf, guard)))
+        },
+    )
+}
+
+pub async fn get_logs(
+    index_scheduler: GuardedData<ActionPolicy<{ actions::METRICS_GET }>, Data<IndexScheduler>>,
+    logs: Data<LogRouteHandle>,
+    body: AwebJson<GetLogs, DeserrJsonError>,
+) -> Result<HttpResponse, ResponseError> {
+    index_scheduler.features().check_logs_route()?;
+
+    let opt = body.into_inner();
+    let mut stream = None;
+
+    logs.modify(|layer| match layer.inner_mut() {
+        None => {
+            // there is no one getting logs
+            *layer.filter_mut() = opt.target.0.clone();
+            let (new_layer, new_stream) = make_layer(&opt, logs.clone());
+
+            *layer.inner_mut() = Some(new_layer);
+            stream = Some(new_stream);
+        }
+        Some(_) => {
+            // there is already someone getting logs
+        }
+    })
+    .unwrap();
+
+    if let Some(stream) = stream {
+        Ok(HttpResponse::Ok().streaming(stream))
+    } else {
+        Err(MeilisearchHttpError::AlreadyUsedLogRoute.into())
+    }
+}
+
+pub async fn cancel_logs(
+    index_scheduler: GuardedData<ActionPolicy<{ actions::METRICS_GET }>, Data<IndexScheduler>>,
+    logs: Data<LogRouteHandle>,
+) -> Result<HttpResponse, ResponseError> {
+    index_scheduler.features().check_logs_route()?;
+
+    if let Err(e) = logs.modify(|layer| *layer.inner_mut() = None) {
+        tracing::error!("Could not free the logs route: {e}");
+    }
+
+    Ok(HttpResponse::NoContent().finish())
+}
+
+#[derive(Debug, Deserr)]
+#[deserr(error = DeserrJsonError, rename_all = camelCase, deny_unknown_fields)]
+pub struct UpdateStderrLogs {
+    #[deserr(default = "info".parse().unwrap(), try_from(&String) = MyTargets::from_str -> DeserrJsonError<BadRequest>)]
+    target: MyTargets,
+}
+
+pub async fn update_stderr_target(
+    index_scheduler: GuardedData<ActionPolicy<{ actions::METRICS_GET }>, Data<IndexScheduler>>,
+    logs: Data<LogStderrHandle>,
+    body: AwebJson<UpdateStderrLogs, DeserrJsonError>,
+) -> Result<HttpResponse, ResponseError> {
+    index_scheduler.features().check_logs_route()?;
+
+    let opt = body.into_inner();
+
+    logs.modify(|layer| {
+        *layer.filter_mut() = opt.target.0.clone();
+    })
+    .unwrap();
+
+    Ok(HttpResponse::NoContent().finish())
+}

crates/meilisearch/src/routes/metrics.rs

@@ -0,0 +1,64 @@
+use actix_web::http::header;
+use actix_web::web::{self, Data};
+use actix_web::HttpResponse;
+use index_scheduler::IndexScheduler;
+use meilisearch_auth::AuthController;
+use meilisearch_types::error::ResponseError;
+use meilisearch_types::keys::actions;
+use prometheus::{Encoder, TextEncoder};
+
+use crate::extractors::authentication::policies::ActionPolicy;
+use crate::extractors::authentication::{AuthenticationError, GuardedData};
+use crate::routes::create_all_stats;
+
+pub fn configure(config: &mut web::ServiceConfig) {
+    config.service(web::resource("").route(web::get().to(get_metrics)));
+}
+
+pub async fn get_metrics(
+    index_scheduler: GuardedData<ActionPolicy<{ actions::METRICS_GET }>, Data<IndexScheduler>>,
+    auth_controller: Data<AuthController>,
+) -> Result<HttpResponse, ResponseError> {
+    index_scheduler.features().check_metrics()?;
+    let auth_filters = index_scheduler.filters();
+    if !auth_filters.all_indexes_authorized() {
+        let mut error = ResponseError::from(AuthenticationError::InvalidToken);
+        error
+            .message
+            .push_str(" The API key for the `/metrics` route must allow access to all indexes.");
+        return Err(error);
+    }
+
+    let response = create_all_stats((*index_scheduler).clone(), auth_controller, auth_filters)?;
+
+    crate::metrics::MEILISEARCH_DB_SIZE_BYTES.set(response.database_size as i64);
+    crate::metrics::MEILISEARCH_USED_DB_SIZE_BYTES.set(response.used_database_size as i64);
+    crate::metrics::MEILISEARCH_INDEX_COUNT.set(response.indexes.len() as i64);
+
+    for (index, value) in response.indexes.iter() {
+        crate::metrics::MEILISEARCH_INDEX_DOCS_COUNT
+            .with_label_values(&[index])
+            .set(value.number_of_documents as i64);
+    }
+
+    for (kind, value) in index_scheduler.get_stats()? {
+        for (value, count) in value {
+            crate::metrics::MEILISEARCH_NB_TASKS
+                .with_label_values(&[&kind, &value])
+                .set(count as i64);
+        }
+    }
+
+    if let Some(last_update) = response.last_update {
+        crate::metrics::MEILISEARCH_LAST_UPDATE.set(last_update.unix_timestamp());
+    }
+    crate::metrics::MEILISEARCH_IS_INDEXING.set(index_scheduler.is_task_processing()? as i64);
+
+    let encoder = TextEncoder::new();
+    let mut buffer = vec![];
+    encoder.encode(&prometheus::gather(), &mut buffer).expect("Failed to encode metrics");
+
+    let response = String::from_utf8(buffer).expect("Failed to convert bytes to string");
+
+    Ok(HttpResponse::Ok().insert_header(header::ContentType(mime::TEXT_PLAIN)).body(response))
+}

crates/meilisearch/src/routes/mod.rs

@@ -0,0 +1,381 @@
+use std::collections::BTreeMap;
+
+use actix_web::web::Data;
+use actix_web::{web, HttpRequest, HttpResponse};
+use index_scheduler::IndexScheduler;
+use meilisearch_auth::AuthController;
+use meilisearch_types::error::{Code, ResponseError};
+use meilisearch_types::settings::{Settings, Unchecked};
+use meilisearch_types::tasks::{Kind, Status, Task, TaskId};
+use serde::{Deserialize, Serialize};
+use time::OffsetDateTime;
+use tracing::debug;
+
+use crate::extractors::authentication::policies::*;
+use crate::extractors::authentication::GuardedData;
+use crate::search_queue::SearchQueue;
+use crate::Opt;
+
+const PAGINATION_DEFAULT_LIMIT: usize = 20;
+
+mod api_key;
+mod dump;
+pub mod features;
+pub mod indexes;
+mod logs;
+mod metrics;
+mod multi_search;
+mod multi_search_analytics;
+mod snapshot;
+mod swap_indexes;
+pub mod tasks;
+
+pub fn configure(cfg: &mut web::ServiceConfig) {
+    cfg.service(web::scope("/tasks").configure(tasks::configure))
+        .service(web::resource("/health").route(web::get().to(get_health)))
+        .service(web::scope("/logs").configure(logs::configure))
+        .service(web::scope("/keys").configure(api_key::configure))
+        .service(web::scope("/dumps").configure(dump::configure))
+        .service(web::scope("/snapshots").configure(snapshot::configure))
+        .service(web::resource("/stats").route(web::get().to(get_stats)))
+        .service(web::resource("/version").route(web::get().to(get_version)))
+        .service(web::scope("/indexes").configure(indexes::configure))
+        .service(web::scope("/multi-search").configure(multi_search::configure))
+        .service(web::scope("/swap-indexes").configure(swap_indexes::configure))
+        .service(web::scope("/metrics").configure(metrics::configure))
+        .service(web::scope("/experimental-features").configure(features::configure));
+}
+
+pub fn get_task_id(req: &HttpRequest, opt: &Opt) -> Result<Option<TaskId>, ResponseError> {
+    if !opt.experimental_replication_parameters {
+        return Ok(None);
+    }
+    let task_id = req
+        .headers()
+        .get("TaskId")
+        .map(|header| {
+            header.to_str().map_err(|e| {
+                ResponseError::from_msg(
+                    format!("TaskId is not a valid utf-8 string: {e}"),
+                    Code::BadRequest,
+                )
+            })
+        })
+        .transpose()?
+        .map(|s| {
+            s.parse::<TaskId>().map_err(|e| {
+                ResponseError::from_msg(
+                    format!(
+                        "Could not parse the TaskId as a {}: {e}",
+                        std::any::type_name::<TaskId>(),
+                    ),
+                    Code::BadRequest,
+                )
+            })
+        })
+        .transpose()?;
+    Ok(task_id)
+}
+
+pub fn is_dry_run(req: &HttpRequest, opt: &Opt) -> Result<bool, ResponseError> {
+    if !opt.experimental_replication_parameters {
+        return Ok(false);
+    }
+    Ok(req
+        .headers()
+        .get("DryRun")
+        .map(|header| {
+            header.to_str().map_err(|e| {
+                ResponseError::from_msg(
+                    format!("DryRun is not a valid utf-8 string: {e}"),
+                    Code::BadRequest,
+                )
+            })
+        })
+        .transpose()?
+        .map_or(false, |s| s.to_lowercase() == "true"))
+}
+
+#[derive(Debug, Serialize)]
+#[serde(rename_all = "camelCase")]
+pub struct SummarizedTaskView {
+    task_uid: TaskId,
+    index_uid: Option<String>,
+    status: Status,
+    #[serde(rename = "type")]
+    kind: Kind,
+    #[serde(serialize_with = "time::serde::rfc3339::serialize")]
+    enqueued_at: OffsetDateTime,
+}
+
+impl From<Task> for SummarizedTaskView {
+    fn from(task: Task) -> Self {
+        SummarizedTaskView {
+            task_uid: task.uid,
+            index_uid: task.index_uid().map(|s| s.to_string()),
+            status: task.status,
+            kind: task.kind.as_kind(),
+            enqueued_at: task.enqueued_at,
+        }
+    }
+}
+
+pub struct Pagination {
+    pub offset: usize,
+    pub limit: usize,
+}
+
+#[derive(Debug, Clone, Serialize)]
+pub struct PaginationView<T> {
+    pub results: Vec<T>,
+    pub offset: usize,
+    pub limit: usize,
+    pub total: usize,
+}
+
+impl Pagination {
+    /// Given the full data to paginate, returns the selected section.
+    pub fn auto_paginate_sized<T>(
+        self,
+        content: impl IntoIterator<Item = T> + ExactSizeIterator,
+    ) -> PaginationView<T>
+    where
+        T: Serialize,
+    {
+        let total = content.len();
+        let content: Vec<_> = content.into_iter().skip(self.offset).take(self.limit).collect();
+        self.format_with(total, content)
+    }
+
+    /// Given an iterator and the total number of elements, returns the selected section.
+    pub fn auto_paginate_unsized<T>(
+        self,
+        total: usize,
+        content: impl IntoIterator<Item = T>,
+    ) -> PaginationView<T>
+    where
+        T: Serialize,
+    {
+        let content: Vec<_> = content.into_iter().skip(self.offset).take(self.limit).collect();
+        self.format_with(total, content)
+    }
+
+    /// Given the data already paginated + the total number of elements, it stores
+    /// everything in a [PaginationResult].
+    pub fn format_with<T>(self, total: usize, results: Vec<T>) -> PaginationView<T>
+    where
+        T: Serialize,
+    {
+        PaginationView { results, offset: self.offset, limit: self.limit, total }
+    }
+}
+
+impl<T> PaginationView<T> {
+    pub fn new(offset: usize, limit: usize, total: usize, results: Vec<T>) -> Self {
+        Self { offset, limit, results, total }
+    }
+}
+
+#[derive(Debug, Clone, Serialize, Deserialize)]
+#[allow(clippy::large_enum_variant)]
+#[serde(tag = "name")]
+pub enum UpdateType {
+    ClearAll,
+    Customs,
+    DocumentsAddition {
+        #[serde(skip_serializing_if = "Option::is_none")]
+        number: Option<usize>,
+    },
+    DocumentsPartial {
+        #[serde(skip_serializing_if = "Option::is_none")]
+        number: Option<usize>,
+    },
+    DocumentsDeletion {
+        #[serde(skip_serializing_if = "Option::is_none")]
+        number: Option<usize>,
+    },
+    Settings {
+        settings: Settings<Unchecked>,
+    },
+}
+
+#[derive(Debug, Clone, Serialize, Deserialize)]
+#[serde(rename_all = "camelCase")]
+pub struct ProcessedUpdateResult {
+    pub update_id: u64,
+    #[serde(rename = "type")]
+    pub update_type: UpdateType,
+    pub duration: f64, // in seconds
+    #[serde(with = "time::serde::rfc3339")]
+    pub enqueued_at: OffsetDateTime,
+    #[serde(with = "time::serde::rfc3339")]
+    pub processed_at: OffsetDateTime,
+}
+
+#[derive(Debug, Clone, Serialize, Deserialize)]
+#[serde(rename_all = "camelCase")]
+pub struct FailedUpdateResult {
+    pub update_id: u64,
+    #[serde(rename = "type")]
+    pub update_type: UpdateType,
+    pub error: ResponseError,
+    pub duration: f64, // in seconds
+    #[serde(with = "time::serde::rfc3339")]
+    pub enqueued_at: OffsetDateTime,
+    #[serde(with = "time::serde::rfc3339")]
+    pub processed_at: OffsetDateTime,
+}
+
+#[derive(Debug, Clone, Serialize, Deserialize)]
+#[serde(rename_all = "camelCase")]
+pub struct EnqueuedUpdateResult {
+    pub update_id: u64,
+    #[serde(rename = "type")]
+    pub update_type: UpdateType,
+    #[serde(with = "time::serde::rfc3339")]
+    pub enqueued_at: OffsetDateTime,
+    #[serde(skip_serializing_if = "Option::is_none", with = "time::serde::rfc3339::option")]
+    pub started_processing_at: Option<OffsetDateTime>,
+}
+
+#[derive(Debug, Clone, Serialize, Deserialize)]
+#[serde(rename_all = "camelCase", tag = "status")]
+pub enum UpdateStatusResponse {
+    Enqueued {
+        #[serde(flatten)]
+        content: EnqueuedUpdateResult,
+    },
+    Processing {
+        #[serde(flatten)]
+        content: EnqueuedUpdateResult,
+    },
+    Failed {
+        #[serde(flatten)]
+        content: FailedUpdateResult,
+    },
+    Processed {
+        #[serde(flatten)]
+        content: ProcessedUpdateResult,
+    },
+}
+
+#[derive(Serialize)]
+#[serde(rename_all = "camelCase")]
+pub struct IndexUpdateResponse {
+    pub update_id: u64,
+}
+
+impl IndexUpdateResponse {
+    pub fn with_id(update_id: u64) -> Self {
+        Self { update_id }
+    }
+}
+
+/// Always return a 200 with:
+/// ```json
+/// {
+///     "status": "Meilisearch is running"
+/// }
+/// ```
+pub async fn running() -> HttpResponse {
+    HttpResponse::Ok().json(serde_json::json!({ "status": "Meilisearch is running" }))
+}
+
+#[derive(Serialize, Debug)]
+#[serde(rename_all = "camelCase")]
+pub struct Stats {
+    pub database_size: u64,
+    #[serde(skip)]
+    pub used_database_size: u64,
+    #[serde(serialize_with = "time::serde::rfc3339::option::serialize")]
+    pub last_update: Option<OffsetDateTime>,
+    pub indexes: BTreeMap<String, indexes::IndexStats>,
+}
+
+async fn get_stats(
+    index_scheduler: GuardedData<ActionPolicy<{ actions::STATS_GET }>, Data<IndexScheduler>>,
+    auth_controller: GuardedData<ActionPolicy<{ actions::STATS_GET }>, Data<AuthController>>,
+) -> Result<HttpResponse, ResponseError> {
+    let filters = index_scheduler.filters();
+
+    let stats = create_all_stats((*index_scheduler).clone(), (*auth_controller).clone(), filters)?;
+
+    debug!(returns = ?stats, "Get stats");
+    Ok(HttpResponse::Ok().json(stats))
+}
+
+pub fn create_all_stats(
+    index_scheduler: Data<IndexScheduler>,
+    auth_controller: Data<AuthController>,
+    filters: &meilisearch_auth::AuthFilter,
+) -> Result<Stats, ResponseError> {
+    let mut last_task: Option<OffsetDateTime> = None;
+    let mut indexes = BTreeMap::new();
+    let mut database_size = 0;
+    let mut used_database_size = 0;
+
+    for index_uid in index_scheduler.index_names()? {
+        // Accumulate the size of all indexes, even unauthorized ones, so
+        // as to return a database_size representative of the correct database size on disk.
+        // See <https://github.com/meilisearch/meilisearch/pull/3541#discussion_r1126747643> for context.
+        let stats = index_scheduler.index_stats(&index_uid)?;
+        database_size += stats.inner_stats.database_size;
+        used_database_size += stats.inner_stats.used_database_size;
+
+        if !filters.is_index_authorized(&index_uid) {
+            continue;
+        }
+
+        last_task = last_task.map_or(Some(stats.inner_stats.updated_at), |last| {
+            Some(last.max(stats.inner_stats.updated_at))
+        });
+        indexes.insert(index_uid.to_string(), stats.into());
+    }
+
+    database_size += index_scheduler.size()?;
+    used_database_size += index_scheduler.used_size()?;
+    database_size += auth_controller.size()?;
+    used_database_size += auth_controller.used_size()?;
+
+    let stats = Stats { database_size, used_database_size, last_update: last_task, indexes };
+    Ok(stats)
+}
+
+#[derive(Serialize)]
+#[serde(rename_all = "camelCase")]
+struct VersionResponse {
+    commit_sha: String,
+    commit_date: String,
+    pkg_version: String,
+}
+
+async fn get_version(
+    _index_scheduler: GuardedData<ActionPolicy<{ actions::VERSION }>, Data<IndexScheduler>>,
+) -> HttpResponse {
+    let build_info = build_info::BuildInfo::from_build();
+
+    HttpResponse::Ok().json(VersionResponse {
+        commit_sha: build_info.commit_sha1.unwrap_or("unknown").to_string(),
+        commit_date: build_info
+            .commit_timestamp
+            .and_then(|commit_timestamp| {
+                commit_timestamp
+                    .format(&time::format_description::well_known::Iso8601::DEFAULT)
+                    .ok()
+            })
+            .unwrap_or("unknown".into()),
+        pkg_version: env!("CARGO_PKG_VERSION").to_string(),
+    })
+}
+
+pub async fn get_health(
+    index_scheduler: Data<IndexScheduler>,
+    auth_controller: Data<AuthController>,
+    search_queue: Data<SearchQueue>,
+) -> Result<HttpResponse, ResponseError> {
+    search_queue.health().unwrap();
+    index_scheduler.health().unwrap();
+    auth_controller.health().unwrap();
+
+    Ok(HttpResponse::Ok().json(serde_json::json!({ "status": "available" })))
+}

crates/meilisearch/src/routes/multi_search.rs

@@ -0,0 +1,185 @@
+use actix_http::StatusCode;
+use actix_web::web::{self, Data};
+use actix_web::{HttpRequest, HttpResponse};
+use deserr::actix_web::AwebJson;
+use index_scheduler::IndexScheduler;
+use meilisearch_types::deserr::DeserrJsonError;
+use meilisearch_types::error::ResponseError;
+use meilisearch_types::keys::actions;
+use serde::Serialize;
+use tracing::debug;
+
+use crate::analytics::Analytics;
+use crate::error::MeilisearchHttpError;
+use crate::extractors::authentication::policies::ActionPolicy;
+use crate::extractors::authentication::{AuthenticationError, GuardedData};
+use crate::extractors::sequential_extractor::SeqHandler;
+use crate::routes::indexes::search::search_kind;
+use crate::search::{
+    add_search_rules, perform_federated_search, perform_search, FederatedSearch, RetrieveVectors,
+    SearchQueryWithIndex, SearchResultWithIndex,
+};
+use crate::search_queue::SearchQueue;
+
+use super::multi_search_analytics::MultiSearchAggregator;
+
+pub fn configure(cfg: &mut web::ServiceConfig) {
+    cfg.service(web::resource("").route(web::post().to(SeqHandler(multi_search_with_post))));
+}
+
+#[derive(Serialize)]
+struct SearchResults {
+    results: Vec<SearchResultWithIndex>,
+}
+
+pub async fn multi_search_with_post(
+    index_scheduler: GuardedData<ActionPolicy<{ actions::SEARCH }>, Data<IndexScheduler>>,
+    search_queue: Data<SearchQueue>,
+    params: AwebJson<FederatedSearch, DeserrJsonError>,
+    req: HttpRequest,
+    analytics: web::Data<Analytics>,
+) -> Result<HttpResponse, ResponseError> {
+    // Since we don't want to process half of the search requests and then get a permit refused
+    // we're going to get one permit for the whole duration of the multi-search request.
+    let permit = search_queue.try_get_search_permit().await?;
+
+    let federated_search = params.into_inner();
+
+    let mut multi_aggregate = MultiSearchAggregator::from_federated_search(&federated_search);
+
+    let FederatedSearch { mut queries, federation } = federated_search;
+
+    let features = index_scheduler.features();
+
+    // regardless of federation, check authorization and apply search rules
+    let auth = 'check_authorization: {
+        for (query_index, federated_query) in queries.iter_mut().enumerate() {
+            let index_uid = federated_query.index_uid.as_str();
+            // Check index from API key
+            if !index_scheduler.filters().is_index_authorized(index_uid) {
+                break 'check_authorization Err(AuthenticationError::InvalidToken)
+                    .with_index(query_index);
+            }
+            // Apply search rules from tenant token
+            if let Some(search_rules) = index_scheduler.filters().get_index_search_rules(index_uid)
+            {
+                add_search_rules(&mut federated_query.filter, search_rules);
+            }
+        }
+        Ok(())
+    };
+
+    auth.map_err(|(mut err, query_index)| {
+        // Add the query index that failed as context for the error message.
+        // We're doing it only here and not directly in the `WithIndex` trait so that the `with_index` function returns a different type
+        // of result and we can benefit from static typing.
+        err.message = format!("Inside `.queries[{query_index}]`: {}", err.message);
+        err
+    })?;
+
+    let response = match federation {
+        Some(federation) => {
+            let search_result = tokio::task::spawn_blocking(move || {
+                perform_federated_search(&index_scheduler, queries, federation, features)
+            })
+            .await;
+            permit.drop().await;
+
+            if let Ok(Ok(_)) = search_result {
+                multi_aggregate.succeed();
+            }
+
+            analytics.publish(multi_aggregate, &req);
+            HttpResponse::Ok().json(search_result??)
+        }
+        None => {
+            // Explicitly expect a `(ResponseError, usize)` for the error type rather than `ResponseError` only,
+            // so that `?` doesn't work if it doesn't use `with_index`, ensuring that it is not forgotten in case of code
+            // changes.
+            let search_results: Result<_, (ResponseError, usize)> = async {
+                let mut search_results = Vec::with_capacity(queries.len());
+                for (query_index, (index_uid, query, federation_options)) in queries
+                    .into_iter()
+                    .map(SearchQueryWithIndex::into_index_query_federation)
+                    .enumerate()
+                {
+                    debug!(on_index = query_index, parameters = ?query, "Multi-search");
+
+                    if federation_options.is_some() {
+                        return Err((
+                            MeilisearchHttpError::FederationOptionsInNonFederatedRequest(
+                                query_index,
+                            )
+                            .into(),
+                            query_index,
+                        ));
+                    }
+
+                    let index = index_scheduler
+                        .index(&index_uid)
+                        .map_err(|err| {
+                            let mut err = ResponseError::from(err);
+                            // Patch the HTTP status code to 400 as it defaults to 404 for `index_not_found`, but
+                            // here the resource not found is not part of the URL.
+                            err.code = StatusCode::BAD_REQUEST;
+                            err
+                        })
+                        .with_index(query_index)?;
+
+                    let search_kind =
+                        search_kind(&query, index_scheduler.get_ref(), &index, features)
+                            .with_index(query_index)?;
+                    let retrieve_vector = RetrieveVectors::new(query.retrieve_vectors, features)
+                        .with_index(query_index)?;
+
+                    let search_result = tokio::task::spawn_blocking(move || {
+                        perform_search(&index, query, search_kind, retrieve_vector, features)
+                    })
+                    .await
+                    .with_index(query_index)?;
+
+                    search_results.push(SearchResultWithIndex {
+                        index_uid: index_uid.into_inner(),
+                        result: search_result.with_index(query_index)?,
+                    });
+                }
+                Ok(search_results)
+            }
+            .await;
+            permit.drop().await;
+
+            if search_results.is_ok() {
+                multi_aggregate.succeed();
+            }
+            analytics.publish(multi_aggregate, &req);
+
+            let search_results = search_results.map_err(|(mut err, query_index)| {
+                // Add the query index that failed as context for the error message.
+                // We're doing it only here and not directly in the `WithIndex` trait so that the `with_index` function returns a different type
+                // of result and we can benefit from static typing.
+                err.message = format!("Inside `.queries[{query_index}]`: {}", err.message);
+                err
+            })?;
+
+            debug!(returns = ?search_results, "Multi-search");
+
+            HttpResponse::Ok().json(SearchResults { results: search_results })
+        }
+    };
+
+    Ok(response)
+}
+
+/// Local `Result` extension trait to avoid `map_err` boilerplate.
+trait WithIndex {
+    type T;
+    /// convert the error type inside of the `Result` to a `ResponseError`, and return a couple of it + the usize.
+    fn with_index(self, index: usize) -> Result<Self::T, (ResponseError, usize)>;
+}
+
+impl<T, E: Into<ResponseError>> WithIndex for Result<T, E> {
+    type T = T;
+    fn with_index(self, index: usize) -> Result<T, (ResponseError, usize)> {
+        self.map_err(|err| (err.into(), index))
+    }
+}

crates/meilisearch/src/routes/multi_search_analytics.rs

@@ -0,0 +1,170 @@
+use std::collections::HashSet;
+
+use serde_json::json;
+
+use crate::{
+    analytics::Aggregate,
+    search::{FederatedSearch, SearchQueryWithIndex},
+};
+
+#[derive(Default)]
+pub struct MultiSearchAggregator {
+    // requests
+    total_received: usize,
+    total_succeeded: usize,
+
+    // sum of the number of distinct indexes in each single request, use with total_received to compute an avg
+    total_distinct_index_count: usize,
+    // number of queries with a single index, use with total_received to compute a proportion
+    total_single_index: usize,
+
+    // sum of the number of search queries in the requests, use with total_received to compute an average
+    total_search_count: usize,
+
+    // scoring
+    show_ranking_score: bool,
+    show_ranking_score_details: bool,
+
+    // federation
+    use_federation: bool,
+}
+
+impl MultiSearchAggregator {
+    pub fn from_federated_search(federated_search: &FederatedSearch) -> Self {
+        let use_federation = federated_search.federation.is_some();
+
+        let distinct_indexes: HashSet<_> = federated_search
+            .queries
+            .iter()
+            .map(|query| {
+                let query = &query;
+                // make sure we get a compilation error if a field gets added to / removed from SearchQueryWithIndex
+                let SearchQueryWithIndex {
+                    index_uid,
+                    federation_options: _,
+                    q: _,
+                    vector: _,
+                    offset: _,
+                    limit: _,
+                    page: _,
+                    hits_per_page: _,
+                    attributes_to_retrieve: _,
+                    retrieve_vectors: _,
+                    attributes_to_crop: _,
+                    crop_length: _,
+                    attributes_to_highlight: _,
+                    show_ranking_score: _,
+                    show_ranking_score_details: _,
+                    show_matches_position: _,
+                    filter: _,
+                    sort: _,
+                    distinct: _,
+                    facets: _,
+                    highlight_pre_tag: _,
+                    highlight_post_tag: _,
+                    crop_marker: _,
+                    matching_strategy: _,
+                    attributes_to_search_on: _,
+                    hybrid: _,
+                    ranking_score_threshold: _,
+                    locales: _,
+                } = query;
+
+                index_uid.as_str()
+            })
+            .collect();
+
+        let show_ranking_score =
+            federated_search.queries.iter().any(|query| query.show_ranking_score);
+        let show_ranking_score_details =
+            federated_search.queries.iter().any(|query| query.show_ranking_score_details);
+
+        Self {
+            total_received: 1,
+            total_succeeded: 0,
+            total_distinct_index_count: distinct_indexes.len(),
+            total_single_index: if distinct_indexes.len() == 1 { 1 } else { 0 },
+            total_search_count: federated_search.queries.len(),
+            show_ranking_score,
+            show_ranking_score_details,
+            use_federation,
+        }
+    }
+
+    pub fn succeed(&mut self) {
+        self.total_succeeded = self.total_succeeded.saturating_add(1);
+    }
+}
+
+impl Aggregate for MultiSearchAggregator {
+    fn event_name(&self) -> &'static str {
+        "Documents Searched by Multi-Search POST"
+    }
+
+    /// Aggregate one [MultiSearchAggregator] into another.
+    fn aggregate(self: Box<Self>, new: Box<Self>) -> Box<Self> {
+        // write the aggregate in a way that will cause a compilation error if a field is added.
+
+        // get ownership of self, replacing it by a default value.
+        let this = *self;
+
+        let total_received = this.total_received.saturating_add(new.total_received);
+        let total_succeeded = this.total_succeeded.saturating_add(new.total_succeeded);
+        let total_distinct_index_count =
+            this.total_distinct_index_count.saturating_add(new.total_distinct_index_count);
+        let total_single_index = this.total_single_index.saturating_add(new.total_single_index);
+        let total_search_count = this.total_search_count.saturating_add(new.total_search_count);
+        let show_ranking_score = this.show_ranking_score || new.show_ranking_score;
+        let show_ranking_score_details =
+            this.show_ranking_score_details || new.show_ranking_score_details;
+        let use_federation = this.use_federation || new.use_federation;
+
+        Box::new(Self {
+            total_received,
+            total_succeeded,
+            total_distinct_index_count,
+            total_single_index,
+            total_search_count,
+            show_ranking_score,
+            show_ranking_score_details,
+            use_federation,
+        })
+    }
+
+    fn into_event(self: Box<Self>) -> serde_json::Value {
+        let Self {
+            total_received,
+            total_succeeded,
+            total_distinct_index_count,
+            total_single_index,
+            total_search_count,
+            show_ranking_score,
+            show_ranking_score_details,
+            use_federation,
+        } = *self;
+
+        json!({
+            "requests": {
+                "total_succeeded": total_succeeded,
+                "total_failed": total_received.saturating_sub(total_succeeded), // just to be sure we never panics
+                "total_received": total_received,
+            },
+            "indexes": {
+                "total_single_index": total_single_index,
+                "total_distinct_index_count": total_distinct_index_count,
+                "avg_distinct_index_count": (total_distinct_index_count as f64) / (total_received as f64), // not 0 else returned early
+            },
+            "searches": {
+                "total_search_count": total_search_count,
+                "avg_search_count": (total_search_count as f64) / (total_received as f64),
+            },
+            "scoring": {
+                "show_ranking_score": show_ranking_score,
+                "show_ranking_score_details": show_ranking_score_details,
+            },
+            "federation": {
+                "use_federation": use_federation,
+            }
+        })
+    }
+}

crates/meilisearch/src/routes/snapshot.rs

@@ -0,0 +1,39 @@
+use actix_web::web::Data;
+use actix_web::{web, HttpRequest, HttpResponse};
+use index_scheduler::IndexScheduler;
+use meilisearch_types::error::ResponseError;
+use meilisearch_types::tasks::KindWithContent;
+use tracing::debug;
+
+use crate::analytics::Analytics;
+use crate::extractors::authentication::policies::*;
+use crate::extractors::authentication::GuardedData;
+use crate::extractors::sequential_extractor::SeqHandler;
+use crate::routes::{get_task_id, is_dry_run, SummarizedTaskView};
+use crate::Opt;
+
+pub fn configure(cfg: &mut web::ServiceConfig) {
+    cfg.service(web::resource("").route(web::post().to(SeqHandler(create_snapshot))));
+}
+
+crate::empty_analytics!(SnapshotAnalytics, "Snapshot Created");
+
+pub async fn create_snapshot(
+    index_scheduler: GuardedData<ActionPolicy<{ actions::SNAPSHOTS_CREATE }>, Data<IndexScheduler>>,
+    req: HttpRequest,
+    opt: web::Data<Opt>,
+    analytics: web::Data<Analytics>,
+) -> Result<HttpResponse, ResponseError> {
+    analytics.publish(SnapshotAnalytics::default(), &req);
+
+    let task = KindWithContent::SnapshotCreation;
+    let uid = get_task_id(&req, &opt)?;
+    let dry_run = is_dry_run(&req, &opt)?;
+    let task: SummarizedTaskView =
+        tokio::task::spawn_blocking(move || index_scheduler.register(task, uid, dry_run))
+            .await??
+            .into();
+
+    debug!(returns = ?task, "Create snapshot");
+    Ok(HttpResponse::Accepted().json(task))
+}

crates/meilisearch/src/routes/swap_indexes.rs

@@ -0,0 +1,87 @@
+use actix_web::web::Data;
+use actix_web::{web, HttpRequest, HttpResponse};
+use deserr::actix_web::AwebJson;
+use deserr::Deserr;
+use index_scheduler::IndexScheduler;
+use meilisearch_types::deserr::DeserrJsonError;
+use meilisearch_types::error::deserr_codes::InvalidSwapIndexes;
+use meilisearch_types::error::ResponseError;
+use meilisearch_types::index_uid::IndexUid;
+use meilisearch_types::tasks::{IndexSwap, KindWithContent};
+use serde::Serialize;
+
+use super::{get_task_id, is_dry_run, SummarizedTaskView};
+use crate::analytics::{Aggregate, Analytics};
+use crate::error::MeilisearchHttpError;
+use crate::extractors::authentication::policies::*;
+use crate::extractors::authentication::{AuthenticationError, GuardedData};
+use crate::extractors::sequential_extractor::SeqHandler;
+use crate::Opt;
+
+pub fn configure(cfg: &mut web::ServiceConfig) {
+    cfg.service(web::resource("").route(web::post().to(SeqHandler(swap_indexes))));
+}
+
+#[derive(Deserr, Debug, Clone, PartialEq, Eq)]
+#[deserr(error = DeserrJsonError, rename_all = camelCase, deny_unknown_fields)]
+pub struct SwapIndexesPayload {
+    #[deserr(error = DeserrJsonError<InvalidSwapIndexes>, missing_field_error = DeserrJsonError::missing_swap_indexes)]
+    indexes: Vec<IndexUid>,
+}
+
+#[derive(Serialize)]
+struct IndexSwappedAnalytics {
+    swap_operation_number: usize,
+}
+
+impl Aggregate for IndexSwappedAnalytics {
+    fn event_name(&self) -> &'static str {
+        "Indexes Swapped"
+    }
+
+    fn aggregate(self: Box<Self>, new: Box<Self>) -> Box<Self> {
+        Box::new(Self {
+            swap_operation_number: self.swap_operation_number.max(new.swap_operation_number),
+        })
+    }
+
+    fn into_event(self: Box<Self>) -> serde_json::Value {
+        serde_json::to_value(*self).unwrap_or_default()
+    }
+}
+
+pub async fn swap_indexes(
+    index_scheduler: GuardedData<ActionPolicy<{ actions::INDEXES_SWAP }>, Data<IndexScheduler>>,
+    params: AwebJson<Vec<SwapIndexesPayload>, DeserrJsonError>,
+    req: HttpRequest,
+    opt: web::Data<Opt>,
+    analytics: web::Data<Analytics>,
+) -> Result<HttpResponse, ResponseError> {
+    let params = params.into_inner();
+    analytics.publish(IndexSwappedAnalytics { swap_operation_number: params.len() }, &req);
+    let filters = index_scheduler.filters();
+
+    let mut swaps = vec![];
+    for SwapIndexesPayload { indexes } in params.into_iter() {
+        // TODO: switch to deserr
+        let (lhs, rhs) = match indexes.as_slice() {
+            [lhs, rhs] => (lhs, rhs),
+            _ => {
+                return Err(MeilisearchHttpError::SwapIndexPayloadWrongLength(indexes).into());
+            }
+        };
+        if !filters.is_index_authorized(lhs) || !filters.is_index_authorized(rhs) {
+            return Err(AuthenticationError::InvalidToken.into());
+        }
+        swaps.push(IndexSwap { indexes: (lhs.to_string(), rhs.to_string()) });
+    }
+
+    let task = KindWithContent::IndexSwap { swaps };
+    let uid = get_task_id(&req, &opt)?;
+    let dry_run = is_dry_run(&req, &opt)?;
+    let task: SummarizedTaskView =
+        tokio::task::spawn_blocking(move || index_scheduler.register(task, uid, dry_run))
+            .await??
+            .into();
+    Ok(HttpResponse::Accepted().json(task))
+}

crates/meilisearch/src/routes/tasks.rs

@@ -0,0 +1,774 @@
+use actix_web::web::Data;
+use actix_web::{web, HttpRequest, HttpResponse};
+use deserr::actix_web::AwebQueryParameter;
+use deserr::Deserr;
+use index_scheduler::{IndexScheduler, Query, TaskId};
+use meilisearch_types::deserr::query_params::Param;
+use meilisearch_types::deserr::DeserrQueryParamError;
+use meilisearch_types::error::deserr_codes::*;
+use meilisearch_types::error::{InvalidTaskDateError, ResponseError};
+use meilisearch_types::index_uid::IndexUid;
+use meilisearch_types::star_or::{OptionStarOr, OptionStarOrList};
+use meilisearch_types::task_view::TaskView;
+use meilisearch_types::tasks::{Kind, KindWithContent, Status};
+use serde::Serialize;
+use time::format_description::well_known::Rfc3339;
+use time::macros::format_description;
+use time::{Date, Duration, OffsetDateTime, Time};
+use tokio::task;
+
+use super::{get_task_id, is_dry_run, SummarizedTaskView};
+use crate::analytics::{Aggregate, AggregateMethod, Analytics};
+use crate::extractors::authentication::policies::*;
+use crate::extractors::authentication::GuardedData;
+use crate::extractors::sequential_extractor::SeqHandler;
+use crate::{aggregate_methods, Opt};
+
+const DEFAULT_LIMIT: u32 = 20;
+
+pub fn configure(cfg: &mut web::ServiceConfig) {
+    cfg.service(
+        web::resource("")
+            .route(web::get().to(SeqHandler(get_tasks)))
+            .route(web::delete().to(SeqHandler(delete_tasks))),
+    )
+    .service(web::resource("/cancel").route(web::post().to(SeqHandler(cancel_tasks))))
+    .service(web::resource("/{task_id}").route(web::get().to(SeqHandler(get_task))));
+}
+#[derive(Debug, Deserr)]
+#[deserr(error = DeserrQueryParamError, rename_all = camelCase, deny_unknown_fields)]
+pub struct TasksFilterQuery {
+    #[deserr(default = Param(DEFAULT_LIMIT), error = DeserrQueryParamError<InvalidTaskLimit>)]
+    pub limit: Param<u32>,
+    #[deserr(default, error = DeserrQueryParamError<InvalidTaskFrom>)]
+    pub from: Option<Param<TaskId>>,
+
+    #[deserr(default, error = DeserrQueryParamError<InvalidTaskUids>)]
+    pub uids: OptionStarOrList<u32>,
+    #[deserr(default, error = DeserrQueryParamError<InvalidTaskCanceledBy>)]
+    pub canceled_by: OptionStarOrList<u32>,
+    #[deserr(default, error = DeserrQueryParamError<InvalidTaskTypes>)]
+    pub types: OptionStarOrList<Kind>,
+    #[deserr(default, error = DeserrQueryParamError<InvalidTaskStatuses>)]
+    pub statuses: OptionStarOrList<Status>,
+    #[deserr(default, error = DeserrQueryParamError<InvalidIndexUid>)]
+    pub index_uids: OptionStarOrList<IndexUid>,
+
+    #[deserr(default, error = DeserrQueryParamError<InvalidTaskAfterEnqueuedAt>, try_from(OptionStarOr<String>) = deserialize_date_after -> InvalidTaskDateError)]
+    pub after_enqueued_at: OptionStarOr<OffsetDateTime>,
+    #[deserr(default, error = DeserrQueryParamError<InvalidTaskBeforeEnqueuedAt>, try_from(OptionStarOr<String>) = deserialize_date_before -> InvalidTaskDateError)]
+    pub before_enqueued_at: OptionStarOr<OffsetDateTime>,
+    #[deserr(default, error = DeserrQueryParamError<InvalidTaskAfterStartedAt>, try_from(OptionStarOr<String>) = deserialize_date_after -> InvalidTaskDateError)]
+    pub after_started_at: OptionStarOr<OffsetDateTime>,
+    #[deserr(default, error = DeserrQueryParamError<InvalidTaskBeforeStartedAt>, try_from(OptionStarOr<String>) = deserialize_date_before -> InvalidTaskDateError)]
+    pub before_started_at: OptionStarOr<OffsetDateTime>,
+    #[deserr(default, error = DeserrQueryParamError<InvalidTaskAfterFinishedAt>, try_from(OptionStarOr<String>) = deserialize_date_after -> InvalidTaskDateError)]
+    pub after_finished_at: OptionStarOr<OffsetDateTime>,
+    #[deserr(default, error = DeserrQueryParamError<InvalidTaskBeforeFinishedAt>, try_from(OptionStarOr<String>) = deserialize_date_before -> InvalidTaskDateError)]
+    pub before_finished_at: OptionStarOr<OffsetDateTime>,
+}
+
+impl TasksFilterQuery {
+    fn into_query(self) -> Query {
+        Query {
+            limit: Some(self.limit.0),
+            from: self.from.as_deref().copied(),
+            statuses: self.statuses.merge_star_and_none(),
+            types: self.types.merge_star_and_none(),
+            index_uids: self.index_uids.map(|x| x.to_string()).merge_star_and_none(),
+            uids: self.uids.merge_star_and_none(),
+            canceled_by: self.canceled_by.merge_star_and_none(),
+            before_enqueued_at: self.before_enqueued_at.merge_star_and_none(),
+            after_enqueued_at: self.after_enqueued_at.merge_star_and_none(),
+            before_started_at: self.before_started_at.merge_star_and_none(),
+            after_started_at: self.after_started_at.merge_star_and_none(),
+            before_finished_at: self.before_finished_at.merge_star_and_none(),
+            after_finished_at: self.after_finished_at.merge_star_and_none(),
+        }
+    }
+}
+
+impl TaskDeletionOrCancelationQuery {
+    fn is_empty(&self) -> bool {
+        matches!(
+            self,
+            TaskDeletionOrCancelationQuery {
+                uids: OptionStarOrList::None,
+                canceled_by: OptionStarOrList::None,
+                types: OptionStarOrList::None,
+                statuses: OptionStarOrList::None,
+                index_uids: OptionStarOrList::None,
+                after_enqueued_at: OptionStarOr::None,
+                before_enqueued_at: OptionStarOr::None,
+                after_started_at: OptionStarOr::None,
+                before_started_at: OptionStarOr::None,
+                after_finished_at: OptionStarOr::None,
+                before_finished_at: OptionStarOr::None
+            }
+        )
+    }
+}
+
+#[derive(Debug, Deserr)]
+#[deserr(error = DeserrQueryParamError, rename_all = camelCase, deny_unknown_fields)]
+pub struct TaskDeletionOrCancelationQuery {
+    #[deserr(default, error = DeserrQueryParamError<InvalidTaskUids>)]
+    pub uids: OptionStarOrList<u32>,
+    #[deserr(default, error = DeserrQueryParamError<InvalidTaskCanceledBy>)]
+    pub canceled_by: OptionStarOrList<u32>,
+    #[deserr(default, error = DeserrQueryParamError<InvalidTaskTypes>)]
+    pub types: OptionStarOrList<Kind>,
+    #[deserr(default, error = DeserrQueryParamError<InvalidTaskStatuses>)]
+    pub statuses: OptionStarOrList<Status>,
+    #[deserr(default, error = DeserrQueryParamError<InvalidIndexUid>)]
+    pub index_uids: OptionStarOrList<IndexUid>,
+
+    #[deserr(default, error = DeserrQueryParamError<InvalidTaskAfterEnqueuedAt>, try_from(OptionStarOr<String>) = deserialize_date_after -> InvalidTaskDateError)]
+    pub after_enqueued_at: OptionStarOr<OffsetDateTime>,
+    #[deserr(default, error = DeserrQueryParamError<InvalidTaskBeforeEnqueuedAt>, try_from(OptionStarOr<String>) = deserialize_date_before -> InvalidTaskDateError)]
+    pub before_enqueued_at: OptionStarOr<OffsetDateTime>,
+    #[deserr(default, error = DeserrQueryParamError<InvalidTaskAfterStartedAt>, try_from(OptionStarOr<String>) = deserialize_date_after -> InvalidTaskDateError)]
+    pub after_started_at: OptionStarOr<OffsetDateTime>,
+    #[deserr(default, error = DeserrQueryParamError<InvalidTaskBeforeStartedAt>, try_from(OptionStarOr<String>) = deserialize_date_before -> InvalidTaskDateError)]
+    pub before_started_at: OptionStarOr<OffsetDateTime>,
+    #[deserr(default, error = DeserrQueryParamError<InvalidTaskAfterFinishedAt>, try_from(OptionStarOr<String>) = deserialize_date_after -> InvalidTaskDateError)]
+    pub after_finished_at: OptionStarOr<OffsetDateTime>,
+    #[deserr(default, error = DeserrQueryParamError<InvalidTaskBeforeFinishedAt>, try_from(OptionStarOr<String>) = deserialize_date_before -> InvalidTaskDateError)]
+    pub before_finished_at: OptionStarOr<OffsetDateTime>,
+}
+
+impl TaskDeletionOrCancelationQuery {
+    fn into_query(self) -> Query {
+        Query {
+            limit: None,
+            from: None,
+            statuses: self.statuses.merge_star_and_none(),
+            types: self.types.merge_star_and_none(),
+            index_uids: self.index_uids.map(|x| x.to_string()).merge_star_and_none(),
+            uids: self.uids.merge_star_and_none(),
+            canceled_by: self.canceled_by.merge_star_and_none(),
+            before_enqueued_at: self.before_enqueued_at.merge_star_and_none(),
+            after_enqueued_at: self.after_enqueued_at.merge_star_and_none(),
+            before_started_at: self.before_started_at.merge_star_and_none(),
+            after_started_at: self.after_started_at.merge_star_and_none(),
+            before_finished_at: self.before_finished_at.merge_star_and_none(),
+            after_finished_at: self.after_finished_at.merge_star_and_none(),
+        }
+    }
+}
+
+aggregate_methods!(
+    CancelTasks => "Tasks Canceled",
+    DeleteTasks => "Tasks Deleted",
+);
+
+#[derive(Serialize)]
+struct TaskFilterAnalytics<Method: AggregateMethod> {
+    filtered_by_uid: bool,
+    filtered_by_index_uid: bool,
+    filtered_by_type: bool,
+    filtered_by_status: bool,
+    filtered_by_canceled_by: bool,
+    filtered_by_before_enqueued_at: bool,
+    filtered_by_after_enqueued_at: bool,
+    filtered_by_before_started_at: bool,
+    filtered_by_after_started_at: bool,
+    filtered_by_before_finished_at: bool,
+    filtered_by_after_finished_at: bool,
+
+    #[serde(skip)]
+    marker: std::marker::PhantomData<Method>,
+}
+
+impl<Method: AggregateMethod + 'static> Aggregate for TaskFilterAnalytics<Method> {
+    fn event_name(&self) -> &'static str {
+        Method::event_name()
+    }
+
+    fn aggregate(self: Box<Self>, new: Box<Self>) -> Box<Self> {
+        Box::new(Self {
+            filtered_by_uid: self.filtered_by_uid | new.filtered_by_uid,
+            filtered_by_index_uid: self.filtered_by_index_uid | new.filtered_by_index_uid,
+            filtered_by_type: self.filtered_by_type | new.filtered_by_type,
+            filtered_by_status: self.filtered_by_status | new.filtered_by_status,
+            filtered_by_canceled_by: self.filtered_by_canceled_by | new.filtered_by_canceled_by,
+            filtered_by_before_enqueued_at: self.filtered_by_before_enqueued_at
+                | new.filtered_by_before_enqueued_at,
+            filtered_by_after_enqueued_at: self.filtered_by_after_enqueued_at
+                | new.filtered_by_after_enqueued_at,
+            filtered_by_before_started_at: self.filtered_by_before_started_at
+                | new.filtered_by_before_started_at,
+            filtered_by_after_started_at: self.filtered_by_after_started_at
+                | new.filtered_by_after_started_at,
+            filtered_by_before_finished_at: self.filtered_by_before_finished_at
+                | new.filtered_by_before_finished_at,
+            filtered_by_after_finished_at: self.filtered_by_after_finished_at
+                | new.filtered_by_after_finished_at,
+
+            marker: std::marker::PhantomData,
+        })
+    }
+
+    fn into_event(self: Box<Self>) -> serde_json::Value {
+        serde_json::to_value(*self).unwrap_or_default()
+    }
+}
+
+async fn cancel_tasks(
+    index_scheduler: GuardedData<ActionPolicy<{ actions::TASKS_CANCEL }>, Data<IndexScheduler>>,
+    params: AwebQueryParameter<TaskDeletionOrCancelationQuery, DeserrQueryParamError>,
+    req: HttpRequest,
+    opt: web::Data<Opt>,
+    analytics: web::Data<Analytics>,
+) -> Result<HttpResponse, ResponseError> {
+    let params = params.into_inner();
+
+    if params.is_empty() {
+        return Err(index_scheduler::Error::TaskCancelationWithEmptyQuery.into());
+    }
+
+    analytics.publish(
+        TaskFilterAnalytics::<CancelTasks> {
+            filtered_by_uid: params.uids.is_some(),
+            filtered_by_index_uid: params.index_uids.is_some(),
+            filtered_by_type: params.types.is_some(),
+            filtered_by_status: params.statuses.is_some(),
+            filtered_by_canceled_by: params.canceled_by.is_some(),
+            filtered_by_before_enqueued_at: params.before_enqueued_at.is_some(),
+            filtered_by_after_enqueued_at: params.after_enqueued_at.is_some(),
+            filtered_by_before_started_at: params.before_started_at.is_some(),
+            filtered_by_after_started_at: params.after_started_at.is_some(),
+            filtered_by_before_finished_at: params.before_finished_at.is_some(),
+            filtered_by_after_finished_at: params.after_finished_at.is_some(),
+
+            marker: std::marker::PhantomData,
+        },
+        &req,
+    );
+
+    let query = params.into_query();
+
+    let (tasks, _) = index_scheduler.get_task_ids_from_authorized_indexes(
+        &index_scheduler.read_txn()?,
+        &query,
+        index_scheduler.filters(),
+    )?;
+    let task_cancelation =
+        KindWithContent::TaskCancelation { query: format!("?{}", req.query_string()), tasks };
+
+    let uid = get_task_id(&req, &opt)?;
+    let dry_run = is_dry_run(&req, &opt)?;
+    let task =
+        task::spawn_blocking(move || index_scheduler.register(task_cancelation, uid, dry_run))
+            .await??;
+    let task: SummarizedTaskView = task.into();
+
+    Ok(HttpResponse::Ok().json(task))
+}
+
+async fn delete_tasks(
+    index_scheduler: GuardedData<ActionPolicy<{ actions::TASKS_DELETE }>, Data<IndexScheduler>>,
+    params: AwebQueryParameter<TaskDeletionOrCancelationQuery, DeserrQueryParamError>,
+    req: HttpRequest,
+    opt: web::Data<Opt>,
+    analytics: web::Data<Analytics>,
+) -> Result<HttpResponse, ResponseError> {
+    let params = params.into_inner();
+
+    if params.is_empty() {
+        return Err(index_scheduler::Error::TaskDeletionWithEmptyQuery.into());
+    }
+
+    analytics.publish(
+        TaskFilterAnalytics::<DeleteTasks> {
+            filtered_by_uid: params.uids.is_some(),
+            filtered_by_index_uid: params.index_uids.is_some(),
+            filtered_by_type: params.types.is_some(),
+            filtered_by_status: params.statuses.is_some(),
+            filtered_by_canceled_by: params.canceled_by.is_some(),
+            filtered_by_before_enqueued_at: params.before_enqueued_at.is_some(),
+            filtered_by_after_enqueued_at: params.after_enqueued_at.is_some(),
+            filtered_by_before_started_at: params.before_started_at.is_some(),
+            filtered_by_after_started_at: params.after_started_at.is_some(),
+            filtered_by_before_finished_at: params.before_finished_at.is_some(),
+            filtered_by_after_finished_at: params.after_finished_at.is_some(),
+
+            marker: std::marker::PhantomData,
+        },
+        &req,
+    );
+
+    let query = params.into_query();
+
+    let (tasks, _) = index_scheduler.get_task_ids_from_authorized_indexes(
+        &index_scheduler.read_txn()?,
+        &query,
+        index_scheduler.filters(),
+    )?;
+    let task_deletion =
+        KindWithContent::TaskDeletion { query: format!("?{}", req.query_string()), tasks };
+
+    let uid = get_task_id(&req, &opt)?;
+    let dry_run = is_dry_run(&req, &opt)?;
+    let task = task::spawn_blocking(move || index_scheduler.register(task_deletion, uid, dry_run))
+        .await??;
+    let task: SummarizedTaskView = task.into();
+
+    Ok(HttpResponse::Ok().json(task))
+}
+
+#[derive(Debug, Serialize)]
+pub struct AllTasks {
+    results: Vec<TaskView>,
+    total: u64,
+    limit: u32,
+    from: Option<u32>,
+    next: Option<u32>,
+}
+
+async fn get_tasks(
+    index_scheduler: GuardedData<ActionPolicy<{ actions::TASKS_GET }>, Data<IndexScheduler>>,
+    params: AwebQueryParameter<TasksFilterQuery, DeserrQueryParamError>,
+) -> Result<HttpResponse, ResponseError> {
+    let mut params = params.into_inner();
+    // We +1 just to know if there is more after this "page" or not.
+    params.limit.0 = params.limit.0.saturating_add(1);
+    let limit = params.limit.0;
+    let query = params.into_query();
+
+    let filters = index_scheduler.filters();
+    let (tasks, total) = index_scheduler.get_tasks_from_authorized_indexes(query, filters)?;
+    let mut results: Vec<_> = tasks.iter().map(TaskView::from_task).collect();
+
+    // If we were able to fetch the number +1 tasks we asked
+    // it means that there is more to come.
+    let next = if results.len() == limit as usize { results.pop().map(|t| t.uid) } else { None };
+
+    let from = results.first().map(|t| t.uid);
+    let tasks = AllTasks { results, limit: limit.saturating_sub(1), total, from, next };
+
+    Ok(HttpResponse::Ok().json(tasks))
+}
+
+async fn get_task(
+    index_scheduler: GuardedData<ActionPolicy<{ actions::TASKS_GET }>, Data<IndexScheduler>>,
+    task_uid: web::Path<String>,
+) -> Result<HttpResponse, ResponseError> {
+    let task_uid_string = task_uid.into_inner();
+
+    let task_uid: TaskId = match task_uid_string.parse() {
+        Ok(id) => id,
+        Err(_e) => {
+            return Err(index_scheduler::Error::InvalidTaskUids { task_uid: task_uid_string }.into())
+        }
+    };
+
+    let query = index_scheduler::Query { uids: Some(vec![task_uid]), ..Query::default() };
+    let filters = index_scheduler.filters();
+    let (tasks, _) = index_scheduler.get_tasks_from_authorized_indexes(query, filters)?;
+
+    if let Some(task) = tasks.first() {
+        let task_view = TaskView::from_task(task);
+        Ok(HttpResponse::Ok().json(task_view))
+    } else {
+        Err(index_scheduler::Error::TaskNotFound(task_uid).into())
+    }
+}
+
+pub enum DeserializeDateOption {
+    Before,
+    After,
+}
+
+pub fn deserialize_date(
+    value: &str,
+    option: DeserializeDateOption,
+) -> std::result::Result<OffsetDateTime, InvalidTaskDateError> {
+    // We can't parse using time's rfc3339 format, since then we won't know what part of the
+    // datetime was not explicitly specified, and thus we won't be able to increment it to the
+    // next step.
+    if let Ok(datetime) = OffsetDateTime::parse(value, &Rfc3339) {
+        // fully specified up to the second
+        // we assume that the subseconds are 0 if not specified, and we don't increment to the next second
+        Ok(datetime)
+    } else if let Ok(datetime) = Date::parse(
+        value,
+        format_description!("[year repr:full base:calendar]-[month repr:numerical]-[day]"),
+    ) {
+        let datetime = datetime.with_time(Time::MIDNIGHT).assume_utc();
+        // add one day since the time was not specified
+        match option {
+            DeserializeDateOption::Before => Ok(datetime),
+            DeserializeDateOption::After => {
+                let datetime = datetime.checked_add(Duration::days(1)).unwrap_or(datetime);
+                Ok(datetime)
+            }
+        }
+    } else {
+        Err(InvalidTaskDateError(value.to_owned()))
+    }
+}
+
+pub fn deserialize_date_after(
+    value: OptionStarOr<String>,
+) -> std::result::Result<OptionStarOr<OffsetDateTime>, InvalidTaskDateError> {
+    value.try_map(|x| deserialize_date(&x, DeserializeDateOption::After))
+}
+pub fn deserialize_date_before(
+    value: OptionStarOr<String>,
+) -> std::result::Result<OptionStarOr<OffsetDateTime>, InvalidTaskDateError> {
+    value.try_map(|x| deserialize_date(&x, DeserializeDateOption::Before))
+}
+
+#[cfg(test)]
+mod tests {
+    use deserr::Deserr;
+    use meili_snap::snapshot;
+    use meilisearch_types::deserr::DeserrQueryParamError;
+    use meilisearch_types::error::{Code, ResponseError};
+
+    use crate::routes::tasks::{TaskDeletionOrCancelationQuery, TasksFilterQuery};
+
+    fn deserr_query_params<T>(j: &str) -> Result<T, ResponseError>
+    where
+        T: Deserr<DeserrQueryParamError>,
+    {
+        let value = serde_urlencoded::from_str::<serde_json::Value>(j)
+            .map_err(|e| ResponseError::from_msg(e.to_string(), Code::BadRequest))?;
+
+        match deserr::deserialize::<_, _, DeserrQueryParamError>(value) {
+            Ok(data) => Ok(data),
+            Err(e) => Err(ResponseError::from(e)),
+        }
+    }
+
+    #[test]
+    fn deserialize_task_filter_dates() {
+        {
+            let params = "afterEnqueuedAt=2021-12-03&beforeEnqueuedAt=2021-12-03&afterStartedAt=2021-12-03&beforeStartedAt=2021-12-03&afterFinishedAt=2021-12-03&beforeFinishedAt=2021-12-03";
+            let query = deserr_query_params::<TaskDeletionOrCancelationQuery>(params).unwrap();
+
+            snapshot!(format!("{:?}", query.after_enqueued_at), @"Other(2021-12-04 0:00:00.0 +00:00:00)");
+            snapshot!(format!("{:?}", query.before_enqueued_at), @"Other(2021-12-03 0:00:00.0 +00:00:00)");
+            snapshot!(format!("{:?}", query.after_started_at), @"Other(2021-12-04 0:00:00.0 +00:00:00)");
+            snapshot!(format!("{:?}", query.before_started_at), @"Other(2021-12-03 0:00:00.0 +00:00:00)");
+            snapshot!(format!("{:?}", query.after_finished_at), @"Other(2021-12-04 0:00:00.0 +00:00:00)");
+            snapshot!(format!("{:?}", query.before_finished_at), @"Other(2021-12-03 0:00:00.0 +00:00:00)");
+        }
+        {
+            let params =
+                "afterEnqueuedAt=2021-12-03T23:45:23Z&beforeEnqueuedAt=2021-12-03T23:45:23Z";
+            let query = deserr_query_params::<TaskDeletionOrCancelationQuery>(params).unwrap();
+            snapshot!(format!("{:?}", query.after_enqueued_at), @"Other(2021-12-03 23:45:23.0 +00:00:00)");
+            snapshot!(format!("{:?}", query.before_enqueued_at), @"Other(2021-12-03 23:45:23.0 +00:00:00)");
+        }
+        {
+            let params = "afterEnqueuedAt=1997-11-12T09:55:06-06:20";
+            let query = deserr_query_params::<TaskDeletionOrCancelationQuery>(params).unwrap();
+            snapshot!(format!("{:?}", query.after_enqueued_at), @"Other(1997-11-12 9:55:06.0 -06:20:00)");
+        }
+        {
+            let params = "afterEnqueuedAt=1997-11-12T09:55:06%2B00:00";
+            let query = deserr_query_params::<TaskDeletionOrCancelationQuery>(params).unwrap();
+            snapshot!(format!("{:?}", query.after_enqueued_at), @"Other(1997-11-12 9:55:06.0 +00:00:00)");
+        }
+        {
+            let params = "afterEnqueuedAt=1997-11-12T09:55:06.200000300Z";
+            let query = deserr_query_params::<TaskDeletionOrCancelationQuery>(params).unwrap();
+            snapshot!(format!("{:?}", query.after_enqueued_at), @"Other(1997-11-12 9:55:06.2000003 +00:00:00)");
+        }
+        {
+            // Stars are allowed in date fields as well
+            let params = "afterEnqueuedAt=*&beforeStartedAt=*&afterFinishedAt=*&beforeFinishedAt=*&afterStartedAt=*&beforeEnqueuedAt=*";
+            let query = deserr_query_params::<TaskDeletionOrCancelationQuery>(params).unwrap();
+            snapshot!(format!("{:?}", query), @"TaskDeletionOrCancelationQuery { uids: None, canceled_by: None, types: None, statuses: None, index_uids: None, after_enqueued_at: Star, before_enqueued_at: Star, after_started_at: Star, before_started_at: Star, after_finished_at: Star, before_finished_at: Star }");
+        }
+        {
+            let params = "afterFinishedAt=2021";
+            let err = deserr_query_params::<TaskDeletionOrCancelationQuery>(params).unwrap_err();
+            snapshot!(meili_snap::json_string!(err), @r###"
+            {
+              "message": "Invalid value in parameter `afterFinishedAt`: `2021` is an invalid date-time. It should follow the YYYY-MM-DD or RFC 3339 date-time format.",
+              "code": "invalid_task_after_finished_at",
+              "type": "invalid_request",
+              "link": "https://docs.meilisearch.com/errors#invalid_task_after_finished_at"
+            }
+            "###);
+        }
+        {
+            let params = "beforeFinishedAt=2021";
+            let err = deserr_query_params::<TaskDeletionOrCancelationQuery>(params).unwrap_err();
+            snapshot!(meili_snap::json_string!(err), @r###"
+            {
+              "message": "Invalid value in parameter `beforeFinishedAt`: `2021` is an invalid date-time. It should follow the YYYY-MM-DD or RFC 3339 date-time format.",
+              "code": "invalid_task_before_finished_at",
+              "type": "invalid_request",
+              "link": "https://docs.meilisearch.com/errors#invalid_task_before_finished_at"
+            }
+            "###);
+        }
+        {
+            let params = "afterEnqueuedAt=2021-12";
+            let err = deserr_query_params::<TaskDeletionOrCancelationQuery>(params).unwrap_err();
+            snapshot!(meili_snap::json_string!(err), @r###"
+            {
+              "message": "Invalid value in parameter `afterEnqueuedAt`: `2021-12` is an invalid date-time. It should follow the YYYY-MM-DD or RFC 3339 date-time format.",
+              "code": "invalid_task_after_enqueued_at",
+              "type": "invalid_request",
+              "link": "https://docs.meilisearch.com/errors#invalid_task_after_enqueued_at"
+            }
+            "###);
+        }
+
+        {
+            let params = "beforeEnqueuedAt=2021-12-03T23";
+            let err = deserr_query_params::<TaskDeletionOrCancelationQuery>(params).unwrap_err();
+            snapshot!(meili_snap::json_string!(err), @r###"
+            {
+              "message": "Invalid value in parameter `beforeEnqueuedAt`: `2021-12-03T23` is an invalid date-time. It should follow the YYYY-MM-DD or RFC 3339 date-time format.",
+              "code": "invalid_task_before_enqueued_at",
+              "type": "invalid_request",
+              "link": "https://docs.meilisearch.com/errors#invalid_task_before_enqueued_at"
+            }
+            "###);
+        }
+        {
+            let params = "afterStartedAt=2021-12-03T23:45";
+            let err = deserr_query_params::<TaskDeletionOrCancelationQuery>(params).unwrap_err();
+            snapshot!(meili_snap::json_string!(err), @r###"
+            {
+              "message": "Invalid value in parameter `afterStartedAt`: `2021-12-03T23:45` is an invalid date-time. It should follow the YYYY-MM-DD or RFC 3339 date-time format.",
+              "code": "invalid_task_after_started_at",
+              "type": "invalid_request",
+              "link": "https://docs.meilisearch.com/errors#invalid_task_after_started_at"
+            }
+            "###);
+        }
+        {
+            let params = "beforeStartedAt=2021-12-03T23:45";
+            let err = deserr_query_params::<TaskDeletionOrCancelationQuery>(params).unwrap_err();
+            snapshot!(meili_snap::json_string!(err), @r###"
+            {
+              "message": "Invalid value in parameter `beforeStartedAt`: `2021-12-03T23:45` is an invalid date-time. It should follow the YYYY-MM-DD or RFC 3339 date-time format.",
+              "code": "invalid_task_before_started_at",
+              "type": "invalid_request",
+              "link": "https://docs.meilisearch.com/errors#invalid_task_before_started_at"
+            }
+            "###);
+        }
+    }
+
+    #[test]
+    fn deserialize_task_filter_uids() {
+        {
+            let params = "uids=78,1,12,73";
+            let query = deserr_query_params::<TaskDeletionOrCancelationQuery>(params).unwrap();
+            snapshot!(format!("{:?}", query.uids), @"List([78, 1, 12, 73])");
+        }
+        {
+            let params = "uids=1";
+            let query = deserr_query_params::<TaskDeletionOrCancelationQuery>(params).unwrap();
+            snapshot!(format!("{:?}", query.uids), @"List([1])");
+        }
+        {
+            let params = "uids=cat,*,dog";
+            let err = deserr_query_params::<TaskDeletionOrCancelationQuery>(params).unwrap_err();
+            snapshot!(meili_snap::json_string!(err), @r###"
+            {
+              "message": "Invalid value in parameter `uids[0]`: could not parse `cat` as a positive integer",
+              "code": "invalid_task_uids",
+              "type": "invalid_request",
+              "link": "https://docs.meilisearch.com/errors#invalid_task_uids"
+            }
+            "###);
+        }
+        {
+            let params = "uids=78,hello,world";
+            let err = deserr_query_params::<TaskDeletionOrCancelationQuery>(params).unwrap_err();
+            snapshot!(meili_snap::json_string!(err), @r###"
+            {
+              "message": "Invalid value in parameter `uids[1]`: could not parse `hello` as a positive integer",
+              "code": "invalid_task_uids",
+              "type": "invalid_request",
+              "link": "https://docs.meilisearch.com/errors#invalid_task_uids"
+            }
+            "###);
+        }
+        {
+            let params = "uids=cat";
+            let err = deserr_query_params::<TaskDeletionOrCancelationQuery>(params).unwrap_err();
+            snapshot!(meili_snap::json_string!(err), @r###"
+            {
+              "message": "Invalid value in parameter `uids`: could not parse `cat` as a positive integer",
+              "code": "invalid_task_uids",
+              "type": "invalid_request",
+              "link": "https://docs.meilisearch.com/errors#invalid_task_uids"
+            }
+            "###);
+        }
+    }
+
+    #[test]
+    fn deserialize_task_filter_status() {
+        {
+            let params = "statuses=succeeded,failed,enqueued,processing,canceled";
+            let query = deserr_query_params::<TaskDeletionOrCancelationQuery>(params).unwrap();
+            snapshot!(format!("{:?}", query.statuses), @"List([Succeeded, Failed, Enqueued, Processing, Canceled])");
+        }
+        {
+            let params = "statuses=enqueued";
+            let query = deserr_query_params::<TaskDeletionOrCancelationQuery>(params).unwrap();
+            snapshot!(format!("{:?}", query.statuses), @"List([Enqueued])");
+        }
+        {
+            let params = "statuses=finished";
+            let err = deserr_query_params::<TaskDeletionOrCancelationQuery>(params).unwrap_err();
+            snapshot!(meili_snap::json_string!(err), @r###"
+            {
+              "message": "Invalid value in parameter `statuses`: `finished` is not a valid task status. Available statuses are `enqueued`, `processing`, `succeeded`, `failed`, `canceled`.",
+              "code": "invalid_task_statuses",
+              "type": "invalid_request",
+              "link": "https://docs.meilisearch.com/errors#invalid_task_statuses"
+            }
+            "###);
+        }
+    }
+    #[test]
+    fn deserialize_task_filter_types() {
+        {
+            let params = "types=documentAdditionOrUpdate,documentDeletion,settingsUpdate,indexCreation,indexDeletion,indexUpdate,indexSwap,taskCancelation,taskDeletion,dumpCreation,snapshotCreation";
+            let query = deserr_query_params::<TaskDeletionOrCancelationQuery>(params).unwrap();
+            snapshot!(format!("{:?}", query.types), @"List([DocumentAdditionOrUpdate, DocumentDeletion, SettingsUpdate, IndexCreation, IndexDeletion, IndexUpdate, IndexSwap, TaskCancelation, TaskDeletion, DumpCreation, SnapshotCreation])");
+        }
+        {
+            let params = "types=settingsUpdate";
+            let query = deserr_query_params::<TaskDeletionOrCancelationQuery>(params).unwrap();
+            snapshot!(format!("{:?}", query.types), @"List([SettingsUpdate])");
+        }
+        {
+            let params = "types=createIndex";
+            let err = deserr_query_params::<TaskDeletionOrCancelationQuery>(params).unwrap_err();
+            snapshot!(meili_snap::json_string!(err), @r###"
+            {
+              "message": "Invalid value in parameter `types`: `createIndex` is not a valid task type. Available types are `documentAdditionOrUpdate`, `documentEdition`, `documentDeletion`, `settingsUpdate`, `indexCreation`, `indexDeletion`, `indexUpdate`, `indexSwap`, `taskCancelation`, `taskDeletion`, `dumpCreation`, `snapshotCreation`.",
+              "code": "invalid_task_types",
+              "type": "invalid_request",
+              "link": "https://docs.meilisearch.com/errors#invalid_task_types"
+            }
+            "###);
+        }
+    }
+    #[test]
+    fn deserialize_task_filter_index_uids() {
+        {
+            let params = "indexUids=toto,tata-78";
+            let query = deserr_query_params::<TaskDeletionOrCancelationQuery>(params).unwrap();
+            snapshot!(format!("{:?}", query.index_uids), @r###"List([IndexUid("toto"), IndexUid("tata-78")])"###);
+        }
+        {
+            let params = "indexUids=index_a";
+            let query = deserr_query_params::<TaskDeletionOrCancelationQuery>(params).unwrap();
+            snapshot!(format!("{:?}", query.index_uids), @r###"List([IndexUid("index_a")])"###);
+        }
+        {
+            let params = "indexUids=1,hé";
+            let err = deserr_query_params::<TaskDeletionOrCancelationQuery>(params).unwrap_err();
+            snapshot!(meili_snap::json_string!(err), @r###"
+            {
+              "message": "Invalid value in parameter `indexUids[1]`: `hé` is not a valid index uid. Index uid can be an integer or a string containing only alphanumeric characters, hyphens (-) and underscores (_), and can not be more than 512 bytes.",
+              "code": "invalid_index_uid",
+              "type": "invalid_request",
+              "link": "https://docs.meilisearch.com/errors#invalid_index_uid"
+            }
+            "###);
+        }
+        {
+            let params = "indexUids=hé";
+            let err = deserr_query_params::<TaskDeletionOrCancelationQuery>(params).unwrap_err();
+            snapshot!(meili_snap::json_string!(err), @r###"
+            {
+              "message": "Invalid value in parameter `indexUids`: `hé` is not a valid index uid. Index uid can be an integer or a string containing only alphanumeric characters, hyphens (-) and underscores (_), and can not be more than 512 bytes.",
+              "code": "invalid_index_uid",
+              "type": "invalid_request",
+              "link": "https://docs.meilisearch.com/errors#invalid_index_uid"
+            }
+            "###);
+        }
+    }
+
+    #[test]
+    fn deserialize_task_filter_general() {
+        {
+            let params = "from=12&limit=15&indexUids=toto,tata-78&statuses=succeeded,enqueued&afterEnqueuedAt=2012-04-23&uids=1,2,3";
+            let query = deserr_query_params::<TasksFilterQuery>(params).unwrap();
+            snapshot!(format!("{:?}", query), @r###"TasksFilterQuery { limit: Param(15), from: Some(Param(12)), uids: List([1, 2, 3]), canceled_by: None, types: None, statuses: List([Succeeded, Enqueued]), index_uids: List([IndexUid("toto"), IndexUid("tata-78")]), after_enqueued_at: Other(2012-04-24 0:00:00.0 +00:00:00), before_enqueued_at: None, after_started_at: None, before_started_at: None, after_finished_at: None, before_finished_at: None }"###);
+        }
+        {
+            // Stars should translate to `None` in the query
+            // Verify value of the default limit
+            let params = "indexUids=*&statuses=succeeded,*&afterEnqueuedAt=2012-04-23&uids=1,2,3";
+            let query = deserr_query_params::<TasksFilterQuery>(params).unwrap();
+            snapshot!(format!("{:?}", query), @"TasksFilterQuery { limit: Param(20), from: None, uids: List([1, 2, 3]), canceled_by: None, types: None, statuses: Star, index_uids: Star, after_enqueued_at: Other(2012-04-24 0:00:00.0 +00:00:00), before_enqueued_at: None, after_started_at: None, before_started_at: None, after_finished_at: None, before_finished_at: None }");
+        }
+        {
+            // Stars should also translate to `None` in task deletion/cancelation queries
+            let params = "indexUids=*&statuses=succeeded,*&afterEnqueuedAt=2012-04-23&uids=1,2,3";
+            let query = deserr_query_params::<TaskDeletionOrCancelationQuery>(params).unwrap();
+            snapshot!(format!("{:?}", query), @"TaskDeletionOrCancelationQuery { uids: List([1, 2, 3]), canceled_by: None, types: None, statuses: Star, index_uids: Star, after_enqueued_at: Other(2012-04-24 0:00:00.0 +00:00:00), before_enqueued_at: None, after_started_at: None, before_started_at: None, after_finished_at: None, before_finished_at: None }");
+        }
+        {
+            // Star in from not allowed
+            let params = "uids=*&from=*";
+            let err = deserr_query_params::<TasksFilterQuery>(params).unwrap_err();
+            snapshot!(meili_snap::json_string!(err), @r###"
+            {
+              "message": "Invalid value in parameter `from`: could not parse `*` as a positive integer",
+              "code": "invalid_task_from",
+              "type": "invalid_request",
+              "link": "https://docs.meilisearch.com/errors#invalid_task_from"
+            }
+            "###);
+        }
+        {
+            // From not allowed in task deletion/cancelation queries
+            let params = "from=12";
+            let err = deserr_query_params::<TaskDeletionOrCancelationQuery>(params).unwrap_err();
+            snapshot!(meili_snap::json_string!(err), @r###"
+            {
+              "message": "Unknown parameter `from`: expected one of `uids`, `canceledBy`, `types`, `statuses`, `indexUids`, `afterEnqueuedAt`, `beforeEnqueuedAt`, `afterStartedAt`, `beforeStartedAt`, `afterFinishedAt`, `beforeFinishedAt`",
+              "code": "bad_request",
+              "type": "invalid_request",
+              "link": "https://docs.meilisearch.com/errors#bad_request"
+            }
+            "###);
+        }
+        {
+            // Limit not allowed in task deletion/cancelation queries
+            let params = "limit=12";
+            let err = deserr_query_params::<TaskDeletionOrCancelationQuery>(params).unwrap_err();
+            snapshot!(meili_snap::json_string!(err), @r###"
+            {
+              "message": "Unknown parameter `limit`: expected one of `uids`, `canceledBy`, `types`, `statuses`, `indexUids`, `afterEnqueuedAt`, `beforeEnqueuedAt`, `afterStartedAt`, `beforeStartedAt`, `afterFinishedAt`, `beforeFinishedAt`",
+              "code": "bad_request",
+              "type": "invalid_request",
+              "link": "https://docs.meilisearch.com/errors#bad_request"
+            }
+            "###);
+        }
+    }
+
+    #[test]
+    fn deserialize_task_delete_or_cancel_empty() {
+        {
+            let params = "";
+            let query = deserr_query_params::<TaskDeletionOrCancelationQuery>(params).unwrap();
+            assert!(query.is_empty());
+        }
+        {
+            let params = "statuses=*";
+            let query = deserr_query_params::<TaskDeletionOrCancelationQuery>(params).unwrap();
+            assert!(!query.is_empty());
+            snapshot!(format!("{query:?}"), @"TaskDeletionOrCancelationQuery { uids: None, canceled_by: None, types: None, statuses: Star, index_uids: None, after_enqueued_at: None, before_enqueued_at: None, after_started_at: None, before_started_at: None, after_finished_at: None, before_finished_at: None }");
+        }
+    }
+}

crates/meilisearch/src/search/federated.rs

@@ -0,0 +1,910 @@
+use std::cmp::Ordering;
+use std::collections::BTreeMap;
+use std::fmt;
+use std::iter::Zip;
+use std::rc::Rc;
+use std::str::FromStr as _;
+use std::time::Duration;
+use std::vec::{IntoIter, Vec};
+
+use actix_http::StatusCode;
+use index_scheduler::{IndexScheduler, RoFeatures};
+use indexmap::IndexMap;
+use meilisearch_types::deserr::DeserrJsonError;
+use meilisearch_types::error::deserr_codes::{
+    InvalidMultiSearchFacetsByIndex, InvalidMultiSearchMaxValuesPerFacet,
+    InvalidMultiSearchMergeFacets, InvalidMultiSearchWeight, InvalidSearchLimit,
+    InvalidSearchOffset,
+};
+use meilisearch_types::error::ResponseError;
+use meilisearch_types::index_uid::IndexUid;
+use meilisearch_types::milli::score_details::{ScoreDetails, ScoreValue};
+use meilisearch_types::milli::{self, DocumentId, OrderBy, TimeBudget};
+use roaring::RoaringBitmap;
+use serde::Serialize;
+
+use super::ranking_rules::{self, RankingRules};
+use super::{
+    compute_facet_distribution_stats, prepare_search, AttributesFormat, ComputedFacets, FacetStats,
+    HitMaker, HitsInfo, RetrieveVectors, SearchHit, SearchKind, SearchQuery, SearchQueryWithIndex,
+};
+use crate::error::MeilisearchHttpError;
+use crate::routes::indexes::search::search_kind;
+
+pub const DEFAULT_FEDERATED_WEIGHT: f64 = 1.0;
+
+#[derive(Debug, Default, Clone, Copy, PartialEq, deserr::Deserr)]
+#[deserr(error = DeserrJsonError, rename_all = camelCase, deny_unknown_fields)]
+pub struct FederationOptions {
+    #[deserr(default, error = DeserrJsonError<InvalidMultiSearchWeight>)]
+    pub weight: Weight,
+}
+
+#[derive(Debug, Clone, Copy, PartialEq, deserr::Deserr)]
+#[deserr(try_from(f64) = TryFrom::try_from -> InvalidMultiSearchWeight)]
+pub struct Weight(f64);
+
+impl Default for Weight {
+    fn default() -> Self {
+        Weight(DEFAULT_FEDERATED_WEIGHT)
+    }
+}
+
+impl std::convert::TryFrom<f64> for Weight {
+    type Error = InvalidMultiSearchWeight;
+
+    fn try_from(f: f64) -> Result<Self, Self::Error> {
+        if f < 0.0 {
+            Err(InvalidMultiSearchWeight)
+        } else {
+            Ok(Weight(f))
+        }
+    }
+}
+
+impl std::ops::Deref for Weight {
+    type Target = f64;
+
+    fn deref(&self) -> &Self::Target {
+        &self.0
+    }
+}
+
+#[derive(Debug, deserr::Deserr)]
+#[deserr(error = DeserrJsonError, rename_all = camelCase, deny_unknown_fields)]
+pub struct Federation {
+    #[deserr(default = super::DEFAULT_SEARCH_LIMIT(), error = DeserrJsonError<InvalidSearchLimit>)]
+    pub limit: usize,
+    #[deserr(default = super::DEFAULT_SEARCH_OFFSET(), error = DeserrJsonError<InvalidSearchOffset>)]
+    pub offset: usize,
+    #[deserr(default, error = DeserrJsonError<InvalidMultiSearchFacetsByIndex>)]
+    pub facets_by_index: BTreeMap<IndexUid, Option<Vec<String>>>,
+    #[deserr(default, error = DeserrJsonError<InvalidMultiSearchMergeFacets>)]
+    pub merge_facets: Option<MergeFacets>,
+}
+
+#[derive(Copy, Clone, Debug, deserr::Deserr, Default)]
+#[deserr(error = DeserrJsonError<InvalidMultiSearchMergeFacets>, rename_all = camelCase, deny_unknown_fields)]
+pub struct MergeFacets {
+    #[deserr(default, error = DeserrJsonError<InvalidMultiSearchMaxValuesPerFacet>)]
+    pub max_values_per_facet: Option<usize>,
+}
+
+#[derive(Debug, deserr::Deserr)]
+#[deserr(error = DeserrJsonError, rename_all = camelCase, deny_unknown_fields)]
+pub struct FederatedSearch {
+    pub queries: Vec<SearchQueryWithIndex>,
+    #[deserr(default)]
+    pub federation: Option<Federation>,
+}
+#[derive(Serialize, Clone)]
+#[serde(rename_all = "camelCase")]
+pub struct FederatedSearchResult {
+    pub hits: Vec<SearchHit>,
+    pub processing_time_ms: u128,
+    #[serde(flatten)]
+    pub hits_info: HitsInfo,
+
+    #[serde(skip_serializing_if = "Option::is_none")]
+    pub semantic_hit_count: Option<u32>,
+
+    #[serde(skip_serializing_if = "Option::is_none")]
+    pub facet_distribution: Option<BTreeMap<String, IndexMap<String, u64>>>,
+    #[serde(skip_serializing_if = "Option::is_none")]
+    pub facet_stats: Option<BTreeMap<String, FacetStats>>,
+    #[serde(skip_serializing_if = "FederatedFacets::is_empty")]
+    pub facets_by_index: FederatedFacets,
+
+    // These fields are only used for analytics purposes
+    #[serde(skip)]
+    pub degraded: bool,
+    #[serde(skip)]
+    pub used_negative_operator: bool,
+}
+
+impl fmt::Debug for FederatedSearchResult {
+    fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
+        let FederatedSearchResult {
+            hits,
+            processing_time_ms,
+            hits_info,
+            semantic_hit_count,
+            degraded,
+            used_negative_operator,
+            facet_distribution,
+            facet_stats,
+            facets_by_index,
+        } = self;
+
+        let mut debug = f.debug_struct("SearchResult");
+        // The most important thing when looking at a search result is the time it took to process
+        debug.field("processing_time_ms", &processing_time_ms);
+        debug.field("hits", &format!("[{} hits returned]", hits.len()));
+        debug.field("hits_info", &hits_info);
+        if *used_negative_operator {
+            debug.field("used_negative_operator", used_negative_operator);
+        }
+        if *degraded {
+            debug.field("degraded", degraded);
+        }
+        if let Some(facet_distribution) = facet_distribution {
+            debug.field("facet_distribution", &facet_distribution);
+        }
+        if let Some(facet_stats) = facet_stats {
+            debug.field("facet_stats", &facet_stats);
+        }
+        if let Some(semantic_hit_count) = semantic_hit_count {
+            debug.field("semantic_hit_count", &semantic_hit_count);
+        }
+        if !facets_by_index.is_empty() {
+            debug.field("facets_by_index", &facets_by_index);
+        }
+
+        debug.finish()
+    }
+}
+
+struct WeightedScore<'a> {
+    details: &'a [ScoreDetails],
+    weight: f64,
+}
+
+impl<'a> WeightedScore<'a> {
+    pub fn new(details: &'a [ScoreDetails], weight: f64) -> Self {
+        Self { details, weight }
+    }
+
+    pub fn weighted_global_score(&self) -> f64 {
+        ScoreDetails::global_score(self.details.iter()) * self.weight
+    }
+
+    pub fn compare_weighted_global_scores(&self, other: &Self) -> Ordering {
+        self.weighted_global_score()
+            .partial_cmp(&other.weighted_global_score())
+            // both are numbers, possibly infinite
+            .unwrap()
+    }
+
+    pub fn compare(&self, other: &Self) -> Ordering {
+        let mut left_it = ScoreDetails::score_values(self.details.iter());
+        let mut right_it = ScoreDetails::score_values(other.details.iter());
+
+        loop {
+            let left = left_it.next();
+            let right = right_it.next();
+
+            match (left, right) {
+                (None, None) => return Ordering::Equal,
+                (None, Some(_)) => return Ordering::Less,
+                (Some(_), None) => return Ordering::Greater,
+                (Some(ScoreValue::Score(left)), Some(ScoreValue::Score(right))) => {
+                    let left = left * self.weight;
+                    let right = right * other.weight;
+                    if (left - right).abs() <= f64::EPSILON {
+                        continue;
+                    }
+                    return left.partial_cmp(&right).unwrap();
+                }
+                (Some(ScoreValue::Sort(left)), Some(ScoreValue::Sort(right))) => {
+                    match left.partial_cmp(right) {
+                        Some(Ordering::Equal) => continue,
+                        Some(order) => return order,
+                        None => return self.compare_weighted_global_scores(other),
+                    }
+                }
+                (Some(ScoreValue::GeoSort(left)), Some(ScoreValue::GeoSort(right))) => {
+                    match left.partial_cmp(right) {
+                        Some(Ordering::Equal) => continue,
+                        Some(order) => return order,
+                        None => {
+                            return self.compare_weighted_global_scores(other);
+                        }
+                    }
+                }
+                // not comparable details, use global
+                (Some(ScoreValue::Score(_)), Some(_))
+                | (Some(_), Some(ScoreValue::Score(_)))
+                | (Some(ScoreValue::GeoSort(_)), Some(ScoreValue::Sort(_)))
+                | (Some(ScoreValue::Sort(_)), Some(ScoreValue::GeoSort(_))) => {
+                    let left_count = left_it.count();
+                    let right_count = right_it.count();
+                    // compare how many remaining groups of rules each side has.
+                    // the group with the most remaining groups wins.
+                    return left_count
+                        .cmp(&right_count)
+                        // breaks ties with the global ranking score
+                        .then_with(|| self.compare_weighted_global_scores(other));
+                }
+            }
+        }
+    }
+}
+
+struct QueryByIndex {
+    query: SearchQuery,
+    federation_options: FederationOptions,
+    query_index: usize,
+}
+
+struct SearchResultByQuery<'a> {
+    documents_ids: Vec<DocumentId>,
+    document_scores: Vec<Vec<ScoreDetails>>,
+    federation_options: FederationOptions,
+    hit_maker: HitMaker<'a>,
+    query_index: usize,
+}
+
+struct SearchResultByQueryIter<'a> {
+    it: Zip<IntoIter<DocumentId>, IntoIter<Vec<ScoreDetails>>>,
+    federation_options: FederationOptions,
+    hit_maker: Rc<HitMaker<'a>>,
+    query_index: usize,
+}
+
+impl<'a> SearchResultByQueryIter<'a> {
+    fn new(
+        SearchResultByQuery {
+            documents_ids,
+            document_scores,
+            federation_options,
+            hit_maker,
+            query_index,
+        }: SearchResultByQuery<'a>,
+    ) -> Self {
+        let it = documents_ids.into_iter().zip(document_scores);
+        Self { it, federation_options, hit_maker: Rc::new(hit_maker), query_index }
+    }
+}
+
+struct SearchResultByQueryIterItem<'a> {
+    docid: DocumentId,
+    score: Vec<ScoreDetails>,
+    federation_options: FederationOptions,
+    hit_maker: Rc<HitMaker<'a>>,
+    query_index: usize,
+}
+
+fn merge_index_local_results(
+    results_by_query: Vec<SearchResultByQuery<'_>>,
+) -> impl Iterator<Item = SearchResultByQueryIterItem> + '_ {
+    itertools::kmerge_by(
+        results_by_query.into_iter().map(SearchResultByQueryIter::new),
+        |left: &SearchResultByQueryIterItem, right: &SearchResultByQueryIterItem| {
+            let left_score = WeightedScore::new(&left.score, *left.federation_options.weight);
+            let right_score = WeightedScore::new(&right.score, *right.federation_options.weight);
+
+            match left_score.compare(&right_score) {
+                // the biggest score goes first
+                Ordering::Greater => true,
+                // break ties using query index
+                Ordering::Equal => left.query_index < right.query_index,
+                Ordering::Less => false,
+            }
+        },
+    )
+}
+
+fn merge_index_global_results(
+    results_by_index: Vec<SearchResultByIndex>,
+) -> impl Iterator<Item = SearchHitByIndex> {
+    itertools::kmerge_by(
+        results_by_index.into_iter().map(|result_by_index| result_by_index.hits.into_iter()),
+        |left: &SearchHitByIndex, right: &SearchHitByIndex| {
+            let left_score = WeightedScore::new(&left.score, *left.federation_options.weight);
+            let right_score = WeightedScore::new(&right.score, *right.federation_options.weight);
+
+            match left_score.compare(&right_score) {
+                // the biggest score goes first
+                Ordering::Greater => true,
+                // break ties using query index
+                Ordering::Equal => left.query_index < right.query_index,
+                Ordering::Less => false,
+            }
+        },
+    )
+}
+
+impl<'a> Iterator for SearchResultByQueryIter<'a> {
+    type Item = SearchResultByQueryIterItem<'a>;
+
+    fn next(&mut self) -> Option<Self::Item> {
+        let (docid, score) = self.it.next()?;
+        Some(SearchResultByQueryIterItem {
+            docid,
+            score,
+            federation_options: self.federation_options,
+            hit_maker: Rc::clone(&self.hit_maker),
+            query_index: self.query_index,
+        })
+    }
+}
+
+struct SearchHitByIndex {
+    hit: SearchHit,
+    score: Vec<ScoreDetails>,
+    federation_options: FederationOptions,
+    query_index: usize,
+}
+
+struct SearchResultByIndex {
+    index: String,
+    hits: Vec<SearchHitByIndex>,
+    estimated_total_hits: usize,
+    degraded: bool,
+    used_negative_operator: bool,
+    facets: Option<ComputedFacets>,
+}
+
+#[derive(Debug, Clone, Default, Serialize)]
+pub struct FederatedFacets(pub BTreeMap<String, ComputedFacets>);
+
+impl FederatedFacets {
+    pub fn insert(&mut self, index: String, facets: Option<ComputedFacets>) {
+        if let Some(facets) = facets {
+            self.0.insert(index, facets);
+        }
+    }
+
+    pub fn is_empty(&self) -> bool {
+        self.0.is_empty()
+    }
+
+    pub fn merge(
+        self,
+        MergeFacets { max_values_per_facet }: MergeFacets,
+        facet_order: BTreeMap<String, (String, OrderBy)>,
+    ) -> Option<ComputedFacets> {
+        if self.is_empty() {
+            return None;
+        }
+
+        let mut distribution: BTreeMap<String, _> = Default::default();
+        let mut stats: BTreeMap<String, FacetStats> = Default::default();
+
+        for facets_by_index in self.0.into_values() {
+            for (facet, index_distribution) in facets_by_index.distribution {
+                match distribution.entry(facet) {
+                    std::collections::btree_map::Entry::Vacant(entry) => {
+                        entry.insert(index_distribution);
+                    }
+                    std::collections::btree_map::Entry::Occupied(mut entry) => {
+                        let distribution = entry.get_mut();
+
+                        for (value, index_count) in index_distribution {
+                            distribution
+                                .entry(value)
+                                .and_modify(|count| *count += index_count)
+                                .or_insert(index_count);
+                        }
+                    }
+                }
+            }
+
+            for (facet, index_stats) in facets_by_index.stats {
+                match stats.entry(facet) {
+                    std::collections::btree_map::Entry::Vacant(entry) => {
+                        entry.insert(index_stats);
+                    }
+                    std::collections::btree_map::Entry::Occupied(mut entry) => {
+                        let stats = entry.get_mut();
+
+                        stats.min = f64::min(stats.min, index_stats.min);
+                        stats.max = f64::max(stats.max, index_stats.max);
+                    }
+                }
+            }
+        }
+
+        // fixup order
+        for (facet, values) in &mut distribution {
+            let order_by = facet_order.get(facet).map(|(_, order)| *order).unwrap_or_default();
+
+            match order_by {
+                OrderBy::Lexicographic => {
+                    values.sort_unstable_by(|left, _, right, _| left.cmp(right))
+                }
+                OrderBy::Count => {
+                    values.sort_unstable_by(|_, left, _, right| {
+                        left.cmp(right)
+                            // biggest first
+                            .reverse()
+                    })
+                }
+            }
+
+            if let Some(max_values_per_facet) = max_values_per_facet {
+                values.truncate(max_values_per_facet)
+            };
+        }
+
+        Some(ComputedFacets { distribution, stats })
+    }
+}
+
+pub fn perform_federated_search(
+    index_scheduler: &IndexScheduler,
+    queries: Vec<SearchQueryWithIndex>,
+    mut federation: Federation,
+    features: RoFeatures,
+) -> Result<FederatedSearchResult, ResponseError> {
+    let before_search = std::time::Instant::now();
+
+    // this implementation partition the queries by index to guarantee an important property:
+    // - all the queries to a particular index use the same read transaction.
+    // This is an important property, otherwise we cannot guarantee the self-consistency of the results.
+
+    // 1. partition queries by index
+    let mut queries_by_index: BTreeMap<String, Vec<QueryByIndex>> = Default::default();
+    for (query_index, federated_query) in queries.into_iter().enumerate() {
+        if let Some(pagination_field) = federated_query.has_pagination() {
+            return Err(MeilisearchHttpError::PaginationInFederatedQuery(
+                query_index,
+                pagination_field,
+            )
+            .into());
+        }
+
+        if let Some(facets) = federated_query.has_facets() {
+            let facets = facets.to_owned();
+            return Err(MeilisearchHttpError::FacetsInFederatedQuery(
+                query_index,
+                federated_query.index_uid.into_inner(),
+                facets,
+            )
+            .into());
+        }
+
+        let (index_uid, query, federation_options) = federated_query.into_index_query_federation();
+
+        queries_by_index.entry(index_uid.into_inner()).or_default().push(QueryByIndex {
+            query,
+            federation_options: federation_options.unwrap_or_default(),
+            query_index,
+        })
+    }
+
+    // 2. perform queries, merge and make hits index by index
+    let required_hit_count = federation.limit + federation.offset;
+
+    // In step (2), semantic_hit_count will be set to Some(0) if any search kind uses semantic
+    // Then in step (3), we'll update its value if there is any semantic search
+    let mut semantic_hit_count = None;
+    let mut results_by_index = Vec::with_capacity(queries_by_index.len());
+    let mut previous_query_data: Option<(RankingRules, usize, String)> = None;
+
+    // remember the order and name of first index for each facet when merging with index settings
+    // to detect if the order is inconsistent for a facet.
+    let mut facet_order: Option<BTreeMap<String, (String, OrderBy)>> = match federation.merge_facets
+    {
+        Some(MergeFacets { .. }) => Some(Default::default()),
+        _ => None,
+    };
+
+    for (index_uid, queries) in queries_by_index {
+        let first_query_index = queries.first().map(|query| query.query_index);
+
+        let index = match index_scheduler.index(&index_uid) {
+            Ok(index) => index,
+            Err(err) => {
+                let mut err = ResponseError::from(err);
+                // Patch the HTTP status code to 400 as it defaults to 404 for `index_not_found`, but
+                // here the resource not found is not part of the URL.
+                err.code = StatusCode::BAD_REQUEST;
+                if let Some(query_index) = first_query_index {
+                    err.message = format!("Inside `.queries[{}]`: {}", query_index, err.message);
+                }
+                return Err(err);
+            }
+        };
+
+        // Important: this is the only transaction we'll use for this index during this federated search
+        let rtxn = index.read_txn()?;
+
+        let criteria = index.criteria(&rtxn)?;
+
+        let dictionary = index.dictionary(&rtxn)?;
+        let dictionary: Option<Vec<_>> =
+            dictionary.as_ref().map(|x| x.iter().map(String::as_str).collect());
+        let separators = index.allowed_separators(&rtxn)?;
+        let separators: Option<Vec<_>> =
+            separators.as_ref().map(|x| x.iter().map(String::as_str).collect());
+
+        // each query gets its individual cutoff
+        let cutoff = index.search_cutoff(&rtxn)?;
+
+        let mut degraded = false;
+        let mut used_negative_operator = false;
+        let mut candidates = RoaringBitmap::new();
+
+        let facets_by_index = federation.facets_by_index.remove(&index_uid).flatten();
+
+        // TODO: recover the max size + facets_by_index as return value of this function so as not to ask it for all queries
+        if let Err(mut error) =
+            check_facet_order(&mut facet_order, &index_uid, &facets_by_index, &index, &rtxn)
+        {
+            error.message = format!(
+                "Inside `.federation.facetsByIndex.{index_uid}`: {error}{}",
+                if let Some(query_index) = first_query_index {
+                    format!("\n - Note: index `{index_uid}` used in `.queries[{query_index}]`")
+                } else {
+                    Default::default()
+                }
+            );
+            return Err(error);
+        }
+
+        // 2.1. Compute all candidates for each query in the index
+        let mut results_by_query = Vec::with_capacity(queries.len());
+
+        for QueryByIndex { query, federation_options, query_index } in queries {
+            // use an immediately invoked lambda to capture the result without returning from the function
+
+            let res: Result<(), ResponseError> = (|| {
+                let search_kind = search_kind(&query, index_scheduler, &index, features)?;
+
+                let canonicalization_kind = match (&search_kind, &query.q) {
+                    (SearchKind::SemanticOnly { .. }, _) => {
+                        ranking_rules::CanonicalizationKind::Vector
+                    }
+                    (_, Some(q)) if !q.is_empty() => ranking_rules::CanonicalizationKind::Keyword,
+                    _ => ranking_rules::CanonicalizationKind::Placeholder,
+                };
+
+                let sort = if let Some(sort) = &query.sort {
+                    let sorts: Vec<_> =
+                        match sort.iter().map(|s| milli::AscDesc::from_str(s)).collect() {
+                            Ok(sorts) => sorts,
+                            Err(asc_desc_error) => {
+                                return Err(milli::Error::from(milli::SortError::from(
+                                    asc_desc_error,
+                                ))
+                                .into())
+                            }
+                        };
+                    Some(sorts)
+                } else {
+                    None
+                };
+
+                let ranking_rules = ranking_rules::RankingRules::new(
+                    criteria.clone(),
+                    sort,
+                    query.matching_strategy.into(),
+                    canonicalization_kind,
+                );
+
+                if let Some((previous_ranking_rules, previous_query_index, previous_index_uid)) =
+                    previous_query_data.take()
+                {
+                    if let Err(error) = ranking_rules.is_compatible_with(&previous_ranking_rules) {
+                        return Err(error.to_response_error(
+                            &ranking_rules,
+                            &previous_ranking_rules,
+                            query_index,
+                            previous_query_index,
+                            &index_uid,
+                            &previous_index_uid,
+                        ));
+                    }
+                    previous_query_data = if previous_ranking_rules.constraint_count()
+                        > ranking_rules.constraint_count()
+                    {
+                        Some((previous_ranking_rules, previous_query_index, previous_index_uid))
+                    } else {
+                        Some((ranking_rules, query_index, index_uid.clone()))
+                    };
+                } else {
+                    previous_query_data = Some((ranking_rules, query_index, index_uid.clone()));
+                }
+
+                match search_kind {
+                    SearchKind::KeywordOnly => {}
+                    _ => semantic_hit_count = Some(0),
+                }
+
+                let retrieve_vectors = RetrieveVectors::new(query.retrieve_vectors, features)?;
+
+                let time_budget = match cutoff {
+                    Some(cutoff) => TimeBudget::new(Duration::from_millis(cutoff)),
+                    None => TimeBudget::default(),
+                };
+
+                let (mut search, _is_finite_pagination, _max_total_hits, _offset) =
+                    prepare_search(&index, &rtxn, &query, &search_kind, time_budget, features)?;
+
+                search.scoring_strategy(milli::score_details::ScoringStrategy::Detailed);
+                search.offset(0);
+                search.limit(required_hit_count);
+
+                let (result, _semantic_hit_count) = super::search_from_kind(search_kind, search)?;
+                let format = AttributesFormat {
+                    attributes_to_retrieve: query.attributes_to_retrieve,
+                    retrieve_vectors,
+                    attributes_to_highlight: query.attributes_to_highlight,
+                    attributes_to_crop: query.attributes_to_crop,
+                    crop_length: query.crop_length,
+                    crop_marker: query.crop_marker,
+                    highlight_pre_tag: query.highlight_pre_tag,
+                    highlight_post_tag: query.highlight_post_tag,
+                    show_matches_position: query.show_matches_position,
+                    sort: query.sort,
+                    show_ranking_score: query.show_ranking_score,
+                    show_ranking_score_details: query.show_ranking_score_details,
+                    locales: query.locales.map(|l| l.iter().copied().map(Into::into).collect()),
+                };
+
+                let milli::SearchResult {
+                    matching_words,
+                    candidates: query_candidates,
+                    documents_ids,
+                    document_scores,
+                    degraded: query_degraded,
+                    used_negative_operator: query_used_negative_operator,
+                } = result;
+
+                candidates |= query_candidates;
+                degraded |= query_degraded;
+                used_negative_operator |= query_used_negative_operator;
+
+                let tokenizer = HitMaker::tokenizer(dictionary.as_deref(), separators.as_deref());
+
+                let formatter_builder = HitMaker::formatter_builder(matching_words, tokenizer);
+
+                let hit_maker = HitMaker::new(&index, &rtxn, format, formatter_builder)?;
+
+                results_by_query.push(SearchResultByQuery {
+                    federation_options,
+                    hit_maker,
+                    query_index,
+                    documents_ids,
+                    document_scores,
+                });
+                Ok(())
+            })();
+
+            if let Err(mut error) = res {
+                error.message = format!("Inside `.queries[{query_index}]`: {}", error.message);
+                return Err(error);
+            }
+        }
+        // 2.2. merge inside index
+        let mut documents_seen = RoaringBitmap::new();
+        let merged_result: Result<Vec<_>, ResponseError> =
+            merge_index_local_results(results_by_query)
+                // skip documents we've already seen & mark that we saw the current document
+                .filter(|SearchResultByQueryIterItem { docid, .. }| documents_seen.insert(*docid))
+                .take(required_hit_count)
+                // 2.3 make hits
+                .map(
+                    |SearchResultByQueryIterItem {
+                         docid,
+                         score,
+                         federation_options,
+                         hit_maker,
+                         query_index,
+                     }| {
+                        let mut hit = hit_maker.make_hit(docid, &score)?;
+                        let weighted_score =
+                            ScoreDetails::global_score(score.iter()) * (*federation_options.weight);
+
+                        let _federation = serde_json::json!(
+                            {
+                                "indexUid": index_uid,
+                                "queriesPosition": query_index,
+                                "weightedRankingScore": weighted_score,
+                            }
+                        );
+                        hit.document.insert("_federation".to_string(), _federation);
+                        Ok(SearchHitByIndex { hit, score, federation_options, query_index })
+                    },
+                )
+                .collect();
+
+        let merged_result = merged_result?;
+
+        let estimated_total_hits = candidates.len() as usize;
+
+        let facets = facets_by_index
+            .map(|facets_by_index| {
+                compute_facet_distribution_stats(
+                    &facets_by_index,
+                    &index,
+                    &rtxn,
+                    candidates,
+                    super::Route::MultiSearch,
+                )
+            })
+            .transpose()
+            .map_err(|mut error| {
+                error.message = format!(
+                    "Inside `.federation.facetsByIndex.{index_uid}`: {}{}",
+                    error.message,
+                    if let Some(query_index) = first_query_index {
+                        format!("\n - Note: index `{index_uid}` used in `.queries[{query_index}]`")
+                    } else {
+                        Default::default()
+                    }
+                );
+                error
+            })?;
+
+        results_by_index.push(SearchResultByIndex {
+            index: index_uid,
+            hits: merged_result,
+            estimated_total_hits,
+            degraded,
+            used_negative_operator,
+            facets,
+        });
+    }
+
+    // bonus step, make sure to return an error if an index wants a non-faceted field, even if no query actually uses that index.
+    for (index_uid, facets) in federation.facets_by_index {
+        let index = match index_scheduler.index(&index_uid) {
+            Ok(index) => index,
+            Err(err) => {
+                let mut err = ResponseError::from(err);
+                // Patch the HTTP status code to 400 as it defaults to 404 for `index_not_found`, but
+                // here the resource not found is not part of the URL.
+                err.code = StatusCode::BAD_REQUEST;
+                err.message = format!(
+                    "Inside `.federation.facetsByIndex.{index_uid}`: {}\n - Note: index `{index_uid}` is not used in queries",
+                    err.message
+                );
+                return Err(err);
+            }
+        };
+
+        // Important: this is the only transaction we'll use for this index during this federated search
+        let rtxn = index.read_txn()?;
+
+        if let Err(mut error) =
+            check_facet_order(&mut facet_order, &index_uid, &facets, &index, &rtxn)
+        {
+            error.message = format!(
+                "Inside `.federation.facetsByIndex.{index_uid}`: {error}\n - Note: index `{index_uid}` is not used in queries",
+            );
+            return Err(error);
+        }
+
+        if let Some(facets) = facets {
+            if let Err(mut error) = compute_facet_distribution_stats(
+                &facets,
+                &index,
+                &rtxn,
+                Default::default(),
+                super::Route::MultiSearch,
+            ) {
+                error.message =
+                    format!("Inside `.federation.facetsByIndex.{index_uid}`: {}\n - Note: index `{index_uid}` is not used in queries", error.message);
+                return Err(error);
+            }
+        }
+    }
+
+    // 3. merge hits and metadata across indexes
+    // 3.1 merge metadata
+    let (estimated_total_hits, degraded, used_negative_operator, facets) = {
+        let mut estimated_total_hits = 0;
+        let mut degraded = false;
+        let mut used_negative_operator = false;
+
+        let mut facets: FederatedFacets = FederatedFacets::default();
+
+        for SearchResultByIndex {
+            index,
+            hits: _,
+            estimated_total_hits: estimated_total_hits_by_index,
+            facets: facets_by_index,
+            degraded: degraded_by_index,
+            used_negative_operator: used_negative_operator_by_index,
+        } in &mut results_by_index
+        {
+            estimated_total_hits += *estimated_total_hits_by_index;
+            degraded |= *degraded_by_index;
+            used_negative_operator |= *used_negative_operator_by_index;
+
+            let facets_by_index = std::mem::take(facets_by_index);
+            let index = std::mem::take(index);
+
+            facets.insert(index, facets_by_index);
+        }
+
+        (estimated_total_hits, degraded, used_negative_operator, facets)
+    };
+
+    // 3.2 merge hits
+    let merged_hits: Vec<_> = merge_index_global_results(results_by_index)
+        .skip(federation.offset)
+        .take(federation.limit)
+        .inspect(|hit| {
+            if let Some(semantic_hit_count) = &mut semantic_hit_count {
+                if hit.score.iter().any(|score| matches!(&score, ScoreDetails::Vector(_))) {
+                    *semantic_hit_count += 1;
+                }
+            }
+        })
+        .map(|hit| hit.hit)
+        .collect();
+
+    let (facet_distribution, facet_stats, facets_by_index) =
+        match federation.merge_facets.zip(facet_order) {
+            Some((merge_facets, facet_order)) => {
+                let facets = facets.merge(merge_facets, facet_order);
+
+                let (facet_distribution, facet_stats) = facets
+                    .map(|ComputedFacets { distribution, stats }| (distribution, stats))
+                    .unzip();
+
+                (facet_distribution, facet_stats, FederatedFacets::default())
+            }
+            None => (None, None, facets),
+        };
+
+    let search_result = FederatedSearchResult {
+        hits: merged_hits,
+        processing_time_ms: before_search.elapsed().as_millis(),
+        hits_info: HitsInfo::OffsetLimit {
+            limit: federation.limit,
+            offset: federation.offset,
+            estimated_total_hits,
+        },
+        semantic_hit_count,
+        degraded,
+        used_negative_operator,
+        facet_distribution,
+        facet_stats,
+        facets_by_index,
+    };
+
+    Ok(search_result)
+}
+
+fn check_facet_order(
+    facet_order: &mut Option<BTreeMap<String, (String, OrderBy)>>,
+    current_index: &str,
+    facets_by_index: &Option<Vec<String>>,
+    index: &milli::Index,
+    rtxn: &milli::heed::RoTxn<'_>,
+) -> Result<(), ResponseError> {
+    if let (Some(facet_order), Some(facets_by_index)) = (facet_order, facets_by_index) {
+        let index_facet_order = index.sort_facet_values_by(rtxn)?;
+        for facet in facets_by_index {
+            let index_facet_order = index_facet_order.get(facet);
+            let (previous_index, previous_facet_order) = facet_order
+                .entry(facet.to_owned())
+                .or_insert_with(|| (current_index.to_owned(), index_facet_order));
+            if previous_facet_order != &index_facet_order {
+                return Err(MeilisearchHttpError::InconsistentFacetOrder {
+                    facet: facet.clone(),
+                    previous_facet_order: *previous_facet_order,
+                    previous_uid: previous_index.clone(),
+                    current_uid: current_index.to_owned(),
+                    index_facet_order,
+                }
+                .into());
+            }
+        }
+    };
+    Ok(())
+}

crates/meilisearch/src/search/ranking_rules.rs

@@ -0,0 +1,823 @@
+use std::collections::HashMap;
+use std::fmt::Write;
+
+use itertools::Itertools as _;
+use meilisearch_types::error::{Code, ResponseError};
+use meilisearch_types::milli::{AscDesc, Criterion, Member, TermsMatchingStrategy};
+
+pub struct RankingRules {
+    canonical_criteria: Vec<Criterion>,
+    canonical_sort: Option<Vec<AscDesc>>,
+    canonicalization_actions: Vec<CanonicalizationAction>,
+    source_criteria: Vec<Criterion>,
+    source_sort: Option<Vec<AscDesc>>,
+}
+
+pub enum CanonicalizationAction {
+    PrependedWords {
+        prepended_index: RankingRuleSource,
+    },
+    RemovedDuplicate {
+        earlier_occurrence: RankingRuleSource,
+        removed_occurrence: RankingRuleSource,
+    },
+    RemovedWords {
+        reason: RemoveWords,
+        removed_occurrence: RankingRuleSource,
+    },
+    RemovedPlaceholder {
+        removed_occurrence: RankingRuleSource,
+    },
+    TruncatedVector {
+        vector_rule: RankingRuleSource,
+        truncated_from: RankingRuleSource,
+    },
+    RemovedVector {
+        vector_rule: RankingRuleSource,
+        removed_occurrence: RankingRuleSource,
+    },
+    RemovedSort {
+        removed_occurrence: RankingRuleSource,
+    },
+}
+
+pub enum RemoveWords {
+    WasPrepended,
+    MatchingStrategyAll,
+}
+
+impl std::fmt::Display for RemoveWords {
+    fn fmt(&self, f: &mut std::fmt::Formatter<'_>) -> std::fmt::Result {
+        let reason = match self {
+            RemoveWords::WasPrepended => "it was previously prepended",
+            RemoveWords::MatchingStrategyAll => "`query.matchingWords` is set to `all`",
+        };
+        f.write_str(reason)
+    }
+}
+
+pub enum CanonicalizationKind {
+    Placeholder,
+    Keyword,
+    Vector,
+}
+
+pub struct CompatibilityError {
+    previous: RankingRule,
+    current: RankingRule,
+}
+impl CompatibilityError {
+    pub(crate) fn to_response_error(
+        &self,
+        ranking_rules: &RankingRules,
+        previous_ranking_rules: &RankingRules,
+        query_index: usize,
+        previous_query_index: usize,
+        index_uid: &str,
+        previous_index_uid: &str,
+    ) -> meilisearch_types::error::ResponseError {
+        let rule = self.current.as_string(
+            &ranking_rules.canonical_criteria,
+            &ranking_rules.canonical_sort,
+            query_index,
+            index_uid,
+        );
+        let previous_rule = self.previous.as_string(
+            &previous_ranking_rules.canonical_criteria,
+            &previous_ranking_rules.canonical_sort,
+            previous_query_index,
+            previous_index_uid,
+        );
+
+        let canonicalization_actions = ranking_rules.canonicalization_notes();
+        let previous_canonicalization_actions = previous_ranking_rules.canonicalization_notes();
+
+        let mut msg = String::new();
+        let reason = self.reason();
+        let _ = writeln!(
+            &mut msg,
+            "The results of queries #{previous_query_index} and #{query_index} are incompatible: "
+        );
+        let _ = writeln!(&mut msg, "  1. {previous_rule}");
+        let _ = writeln!(&mut msg, "  2. {rule}");
+        let _ = writeln!(&mut msg, "  - {reason}");
+
+        if !previous_canonicalization_actions.is_empty() {
+            let _ = write!(&mut msg, "  - note: The ranking rules of query #{previous_query_index} were modified during canonicalization:\n{previous_canonicalization_actions}");
+        }
+
+        if !canonicalization_actions.is_empty() {
+            let _ = write!(&mut msg, "  - note: The ranking rules of query #{query_index} were modified during canonicalization:\n{canonicalization_actions}");
+        }
+
+        ResponseError::from_msg(msg, Code::InvalidMultiSearchQueryRankingRules)
+    }
+    pub fn reason(&self) -> &'static str {
+        match (self.previous.kind, self.current.kind) {
+            (RankingRuleKind::Relevancy, RankingRuleKind::AscendingSort)
+            | (RankingRuleKind::Relevancy, RankingRuleKind::DescendingSort)
+            | (RankingRuleKind::AscendingSort, RankingRuleKind::Relevancy)
+            | (RankingRuleKind::DescendingSort, RankingRuleKind::Relevancy) => {
+                "cannot compare a relevancy rule with a sort rule"
+            }
+
+            (RankingRuleKind::Relevancy, RankingRuleKind::AscendingGeoSort)
+            | (RankingRuleKind::Relevancy, RankingRuleKind::DescendingGeoSort)
+            | (RankingRuleKind::AscendingGeoSort, RankingRuleKind::Relevancy)
+            | (RankingRuleKind::DescendingGeoSort, RankingRuleKind::Relevancy) => {
+                "cannot compare a relevancy rule with a geosort rule"
+            }
+
+            (RankingRuleKind::AscendingSort, RankingRuleKind::DescendingSort)
+            | (RankingRuleKind::DescendingSort, RankingRuleKind::AscendingSort) => {
+                "cannot compare two sort rules in opposite directions"
+            }
+
+            (RankingRuleKind::AscendingSort, RankingRuleKind::AscendingGeoSort)
+            | (RankingRuleKind::AscendingSort, RankingRuleKind::DescendingGeoSort)
+            | (RankingRuleKind::DescendingSort, RankingRuleKind::AscendingGeoSort)
+            | (RankingRuleKind::DescendingSort, RankingRuleKind::DescendingGeoSort)
+            | (RankingRuleKind::AscendingGeoSort, RankingRuleKind::AscendingSort)
+            | (RankingRuleKind::AscendingGeoSort, RankingRuleKind::DescendingSort)
+            | (RankingRuleKind::DescendingGeoSort, RankingRuleKind::AscendingSort)
+            | (RankingRuleKind::DescendingGeoSort, RankingRuleKind::DescendingSort) => {
+                "cannot compare a sort rule with a geosort rule"
+            }
+
+            (RankingRuleKind::AscendingGeoSort, RankingRuleKind::DescendingGeoSort)
+            | (RankingRuleKind::DescendingGeoSort, RankingRuleKind::AscendingGeoSort) => {
+                "cannot compare two geosort rules in opposite directions"
+            }
+            (RankingRuleKind::Relevancy, RankingRuleKind::Relevancy)
+            | (RankingRuleKind::AscendingSort, RankingRuleKind::AscendingSort)
+            | (RankingRuleKind::DescendingSort, RankingRuleKind::DescendingSort)
+            | (RankingRuleKind::AscendingGeoSort, RankingRuleKind::AscendingGeoSort)
+            | (RankingRuleKind::DescendingGeoSort, RankingRuleKind::DescendingGeoSort) => {
+                "internal error, comparison should be possible"
+            }
+        }
+    }
+}
+
+impl RankingRules {
+    pub fn new(
+        criteria: Vec<Criterion>,
+        sort: Option<Vec<AscDesc>>,
+        terms_matching_strategy: TermsMatchingStrategy,
+        canonicalization_kind: CanonicalizationKind,
+    ) -> Self {
+        let (canonical_criteria, canonical_sort, canonicalization_actions) =
+            Self::canonicalize(&criteria, &sort, terms_matching_strategy, canonicalization_kind);
+        Self {
+            canonical_criteria,
+            canonical_sort,
+            canonicalization_actions,
+            source_criteria: criteria,
+            source_sort: sort,
+        }
+    }
+
+    fn canonicalize(
+        criteria: &[Criterion],
+        sort: &Option<Vec<AscDesc>>,
+        terms_matching_strategy: TermsMatchingStrategy,
+        canonicalization_kind: CanonicalizationKind,
+    ) -> (Vec<Criterion>, Option<Vec<AscDesc>>, Vec<CanonicalizationAction>) {
+        match canonicalization_kind {
+            CanonicalizationKind::Placeholder => Self::canonicalize_placeholder(criteria, sort),
+            CanonicalizationKind::Keyword => {
+                Self::canonicalize_keyword(criteria, sort, terms_matching_strategy)
+            }
+            CanonicalizationKind::Vector => Self::canonicalize_vector(criteria, sort),
+        }
+    }
+
+    fn canonicalize_placeholder(
+        criteria: &[Criterion],
+        sort_query: &Option<Vec<AscDesc>>,
+    ) -> (Vec<Criterion>, Option<Vec<AscDesc>>, Vec<CanonicalizationAction>) {
+        let mut sort = None;
+
+        let mut sorted_fields = HashMap::new();
+        let mut canonicalization_actions = Vec::new();
+        let mut canonical_criteria = Vec::new();
+        let mut canonical_sort = None;
+
+        for (criterion_index, criterion) in criteria.iter().enumerate() {
+            match criterion.clone() {
+                Criterion::Words
+                | Criterion::Typo
+                | Criterion::Proximity
+                | Criterion::Attribute
+                | Criterion::Exactness => {
+                    canonicalization_actions.push(CanonicalizationAction::RemovedPlaceholder {
+                        removed_occurrence: RankingRuleSource::Criterion(criterion_index),
+                    })
+                }
+
+                Criterion::Sort => {
+                    if let Some(previous_index) = sort {
+                        canonicalization_actions.push(CanonicalizationAction::RemovedDuplicate {
+                            earlier_occurrence: RankingRuleSource::Criterion(previous_index),
+                            removed_occurrence: RankingRuleSource::Criterion(criterion_index),
+                        });
+                    } else if let Some(sort_query) = sort_query {
+                        sort = Some(criterion_index);
+                        canonical_criteria.push(criterion.clone());
+                        canonical_sort = Some(canonicalize_sort(
+                            &mut sorted_fields,
+                            sort_query.as_slice(),
+                            criterion_index,
+                            &mut canonicalization_actions,
+                        ));
+                    } else {
+                        canonicalization_actions.push(CanonicalizationAction::RemovedSort {
+                            removed_occurrence: RankingRuleSource::Criterion(criterion_index),
+                        })
+                    }
+                }
+                Criterion::Asc(s) | Criterion::Desc(s) => match sorted_fields.entry(s) {
+                    std::collections::hash_map::Entry::Occupied(entry) => canonicalization_actions
+                        .push(CanonicalizationAction::RemovedDuplicate {
+                            earlier_occurrence: *entry.get(),
+                            removed_occurrence: RankingRuleSource::Criterion(criterion_index),
+                        }),
+                    std::collections::hash_map::Entry::Vacant(entry) => {
+                        entry.insert(RankingRuleSource::Criterion(criterion_index));
+                        canonical_criteria.push(criterion.clone())
+                    }
+                },
+            }
+        }
+
+        (canonical_criteria, canonical_sort, canonicalization_actions)
+    }
+
+    fn canonicalize_vector(
+        criteria: &[Criterion],
+        sort_query: &Option<Vec<AscDesc>>,
+    ) -> (Vec<Criterion>, Option<Vec<AscDesc>>, Vec<CanonicalizationAction>) {
+        let mut sort = None;
+
+        let mut sorted_fields = HashMap::new();
+        let mut canonicalization_actions = Vec::new();
+        let mut canonical_criteria = Vec::new();
+        let mut canonical_sort = None;
+
+        let mut vector = None;
+
+        'criteria: for (criterion_index, criterion) in criteria.iter().enumerate() {
+            match criterion.clone() {
+                Criterion::Words
+                | Criterion::Typo
+                | Criterion::Proximity
+                | Criterion::Attribute
+                | Criterion::Exactness => match vector {
+                    Some(previous_occurrence) => {
+                        if sorted_fields.is_empty() {
+                            canonicalization_actions.push(CanonicalizationAction::RemovedVector {
+                                vector_rule: RankingRuleSource::Criterion(previous_occurrence),
+                                removed_occurrence: RankingRuleSource::Criterion(criterion_index),
+                            });
+                        } else {
+                            canonicalization_actions.push(
+                                CanonicalizationAction::TruncatedVector {
+                                    vector_rule: RankingRuleSource::Criterion(previous_occurrence),
+                                    truncated_from: RankingRuleSource::Criterion(criterion_index),
+                                },
+                            );
+                            break 'criteria;
+                        }
+                    }
+                    None => {
+                        canonical_criteria.push(criterion.clone());
+                        vector = Some(criterion_index);
+                    }
+                },
+
+                Criterion::Sort => {
+                    if let Some(previous_index) = sort {
+                        canonicalization_actions.push(CanonicalizationAction::RemovedDuplicate {
+                            earlier_occurrence: RankingRuleSource::Criterion(previous_index),
+                            removed_occurrence: RankingRuleSource::Criterion(criterion_index),
+                        });
+                    } else if let Some(sort_query) = sort_query {
+                        sort = Some(criterion_index);
+                        canonical_criteria.push(criterion.clone());
+                        canonical_sort = Some(canonicalize_sort(
+                            &mut sorted_fields,
+                            sort_query.as_slice(),
+                            criterion_index,
+                            &mut canonicalization_actions,
+                        ));
+                    } else {
+                        canonicalization_actions.push(CanonicalizationAction::RemovedSort {
+                            removed_occurrence: RankingRuleSource::Criterion(criterion_index),
+                        })
+                    }
+                }
+                Criterion::Asc(s) | Criterion::Desc(s) => match sorted_fields.entry(s) {
+                    std::collections::hash_map::Entry::Occupied(entry) => canonicalization_actions
+                        .push(CanonicalizationAction::RemovedDuplicate {
+                            earlier_occurrence: *entry.get(),
+                            removed_occurrence: RankingRuleSource::Criterion(criterion_index),
+                        }),
+                    std::collections::hash_map::Entry::Vacant(entry) => {
+                        entry.insert(RankingRuleSource::Criterion(criterion_index));
+                        canonical_criteria.push(criterion.clone())
+                    }
+                },
+            }
+        }
+
+        (canonical_criteria, canonical_sort, canonicalization_actions)
+    }
+
+    fn canonicalize_keyword(
+        criteria: &[Criterion],
+        sort_query: &Option<Vec<AscDesc>>,
+        terms_matching_strategy: TermsMatchingStrategy,
+    ) -> (Vec<Criterion>, Option<Vec<AscDesc>>, Vec<CanonicalizationAction>) {
+        let mut words = None;
+        let mut typo = None;
+        let mut proximity = None;
+        let mut sort = None;
+        let mut attribute = None;
+        let mut exactness = None;
+        let mut sorted_fields = HashMap::new();
+
+        let mut canonical_criteria = Vec::new();
+        let mut canonical_sort = None;
+
+        let mut canonicalization_actions = Vec::new();
+
+        for (criterion_index, criterion) in criteria.iter().enumerate() {
+            let criterion = criterion.clone();
+            match criterion.clone() {
+                Criterion::Words => {
+                    if let TermsMatchingStrategy::All = terms_matching_strategy {
+                        canonicalization_actions.push(CanonicalizationAction::RemovedWords {
+                            reason: RemoveWords::MatchingStrategyAll,
+                            removed_occurrence: RankingRuleSource::Criterion(criterion_index),
+                        });
+                        continue;
+                    }
+                    if let Some(maybe_previous_index) = words {
+                        if let Some(previous_index) = maybe_previous_index {
+                            canonicalization_actions.push(
+                                CanonicalizationAction::RemovedDuplicate {
+                                    earlier_occurrence: RankingRuleSource::Criterion(
+                                        previous_index,
+                                    ),
+                                    removed_occurrence: RankingRuleSource::Criterion(
+                                        criterion_index,
+                                    ),
+                                },
+                            );
+                            continue;
+                        }
+                        canonicalization_actions.push(CanonicalizationAction::RemovedWords {
+                            reason: RemoveWords::WasPrepended,
+                            removed_occurrence: RankingRuleSource::Criterion(criterion_index),
+                        })
+                    }
+                    words = Some(Some(criterion_index));
+                    canonical_criteria.push(criterion);
+                }
+                Criterion::Typo => {
+                    canonicalize_criterion(
+                        criterion,
+                        criterion_index,
+                        terms_matching_strategy,
+                        &mut words,
+                        &mut canonicalization_actions,
+                        &mut canonical_criteria,
+                        &mut typo,
+                    );
+                }
+                Criterion::Proximity => {
+                    canonicalize_criterion(
+                        criterion,
+                        criterion_index,
+                        terms_matching_strategy,
+                        &mut words,
+                        &mut canonicalization_actions,
+                        &mut canonical_criteria,
+                        &mut proximity,
+                    );
+                }
+                Criterion::Attribute => {
+                    canonicalize_criterion(
+                        criterion,
+                        criterion_index,
+                        terms_matching_strategy,
+                        &mut words,
+                        &mut canonicalization_actions,
+                        &mut canonical_criteria,
+                        &mut attribute,
+                    );
+                }
+                Criterion::Exactness => {
+                    canonicalize_criterion(
+                        criterion,
+                        criterion_index,
+                        terms_matching_strategy,
+                        &mut words,
+                        &mut canonicalization_actions,
+                        &mut canonical_criteria,
+                        &mut exactness,
+                    );
+                }
+
+                Criterion::Sort => {
+                    if let Some(previous_index) = sort {
+                        canonicalization_actions.push(CanonicalizationAction::RemovedDuplicate {
+                            earlier_occurrence: RankingRuleSource::Criterion(previous_index),
+                            removed_occurrence: RankingRuleSource::Criterion(criterion_index),
+                        });
+                    } else if let Some(sort_query) = sort_query {
+                        sort = Some(criterion_index);
+                        canonical_criteria.push(criterion);
+                        canonical_sort = Some(canonicalize_sort(
+                            &mut sorted_fields,
+                            sort_query.as_slice(),
+                            criterion_index,
+                            &mut canonicalization_actions,
+                        ));
+                    } else {
+                        canonicalization_actions.push(CanonicalizationAction::RemovedSort {
+                            removed_occurrence: RankingRuleSource::Criterion(criterion_index),
+                        })
+                    }
+                }
+                Criterion::Asc(s) | Criterion::Desc(s) => match sorted_fields.entry(s) {
+                    std::collections::hash_map::Entry::Occupied(entry) => canonicalization_actions
+                        .push(CanonicalizationAction::RemovedDuplicate {
+                            earlier_occurrence: *entry.get(),
+                            removed_occurrence: RankingRuleSource::Criterion(criterion_index),
+                        }),
+                    std::collections::hash_map::Entry::Vacant(entry) => {
+                        entry.insert(RankingRuleSource::Criterion(criterion_index));
+                        canonical_criteria.push(criterion)
+                    }
+                },
+            }
+        }
+
+        (canonical_criteria, canonical_sort, canonicalization_actions)
+    }
+
+    pub fn is_compatible_with(&self, previous: &Self) -> Result<(), CompatibilityError> {
+        for (current, previous) in self.coalesce_iterator().zip(previous.coalesce_iterator()) {
+            if current.kind != previous.kind {
+                return Err(CompatibilityError { current, previous });
+            }
+        }
+        Ok(())
+    }
+
+    pub fn constraint_count(&self) -> usize {
+        self.coalesce_iterator().count()
+    }
+
+    fn coalesce_iterator(&self) -> impl Iterator<Item = RankingRule> + '_ {
+        self.canonical_criteria
+            .iter()
+            .enumerate()
+            .flat_map(|(criterion_index, criterion)| {
+                RankingRule::from_criterion(criterion_index, criterion, &self.canonical_sort)
+            })
+            .coalesce(
+                |previous @ RankingRule { source: previous_source, kind: previous_kind },
+                 current @ RankingRule { source, kind }| {
+                    match (previous_kind, kind) {
+                        (RankingRuleKind::Relevancy, RankingRuleKind::Relevancy) => {
+                            let merged_source = match (previous_source, source) {
+                                (
+                                    RankingRuleSource::Criterion(previous),
+                                    RankingRuleSource::Criterion(current),
+                                ) => RankingRuleSource::CoalescedCriteria(previous, current),
+                                (
+                                    RankingRuleSource::CoalescedCriteria(begin, _end),
+                                    RankingRuleSource::Criterion(current),
+                                ) => RankingRuleSource::CoalescedCriteria(begin, current),
+                                (_previous, current) => current,
+                            };
+                            Ok(RankingRule { source: merged_source, kind })
+                        }
+                        _ => Err((previous, current)),
+                    }
+                },
+            )
+    }
+
+    fn canonicalization_notes(&self) -> String {
+        use CanonicalizationAction::*;
+        let mut notes = String::new();
+        for (index, action) in self.canonicalization_actions.iter().enumerate() {
+            let index = index + 1;
+            let _ = match action {
+                PrependedWords { prepended_index } => writeln!(
+                    &mut notes,
+                    "    {index}. Prepended rule `words` before first relevancy rule `{}` at position {}",
+                    prepended_index.rule_name(&self.source_criteria, &self.source_sort),
+                    prepended_index.rule_position()
+                ),
+                RemovedDuplicate { earlier_occurrence, removed_occurrence } => writeln!(
+                    &mut notes,
+                    "    {index}. Removed duplicate rule `{}` at position {} as it already appears at position {}",
+                    earlier_occurrence.rule_name(&self.source_criteria, &self.source_sort),
+                    removed_occurrence.rule_position(),
+                    earlier_occurrence.rule_position(),
+                ),
+                RemovedWords { reason, removed_occurrence } => writeln!(
+                    &mut notes,
+                    "    {index}. Removed rule `words` at position {} because {reason}",
+                    removed_occurrence.rule_position()
+                ),
+                RemovedPlaceholder { removed_occurrence } => writeln!(
+                    &mut notes,
+                    "    {index}. Removed relevancy rule `{}` at position {} because the query is a placeholder search (`q`: \"\")",
+                    removed_occurrence.rule_name(&self.source_criteria, &self.source_sort),
+                    removed_occurrence.rule_position()
+                ),
+                TruncatedVector { vector_rule, truncated_from } => writeln!(
+                    &mut notes,
+                    "    {index}. Truncated relevancy rule `{}` at position {} and later rules because the query is a vector search and `vector` was inserted at position {}",
+                    truncated_from.rule_name(&self.source_criteria, &self.source_sort),
+                    truncated_from.rule_position(),
+                    vector_rule.rule_position(),
+                ),
+                RemovedVector { vector_rule, removed_occurrence } => writeln!(
+                    &mut notes,
+                    "    {index}. Removed relevancy rule `{}` at position {} because the query is a vector search and `vector` was already inserted at position {}",
+                    removed_occurrence.rule_name(&self.source_criteria, &self.source_sort),
+                    removed_occurrence.rule_position(),
+                    vector_rule.rule_position(),
+                ),
+                RemovedSort { removed_occurrence } => writeln!(
+                    &mut notes,
+                    "   {index}. Removed rule `sort` at position {} because `query.sort` is empty",
+removed_occurrence.rule_position()
+                ),
+            };
+        }
+        notes
+    }
+}
+
+fn canonicalize_sort(
+    sorted_fields: &mut HashMap<String, RankingRuleSource>,
+    sort_query: &[AscDesc],
+    criterion_index: usize,
+    canonicalization_actions: &mut Vec<CanonicalizationAction>,
+) -> Vec<AscDesc> {
+    let mut geo_sorted = None;
+    let mut canonical_sort = Vec::new();
+    for (sort_index, asc_desc) in sort_query.iter().enumerate() {
+        let source = RankingRuleSource::Sort { criterion_index, sort_index };
+        let asc_desc = asc_desc.clone();
+        match asc_desc.clone() {
+            AscDesc::Asc(Member::Field(s)) | AscDesc::Desc(Member::Field(s)) => {
+                match sorted_fields.entry(s) {
+                    std::collections::hash_map::Entry::Occupied(entry) => canonicalization_actions
+                        .push(CanonicalizationAction::RemovedDuplicate {
+                            earlier_occurrence: *entry.get(),
+                            removed_occurrence: source,
+                        }),
+                    std::collections::hash_map::Entry::Vacant(entry) => {
+                        entry.insert(source);
+                        canonical_sort.push(asc_desc);
+                    }
+                }
+            }
+            AscDesc::Asc(Member::Geo(_)) | AscDesc::Desc(Member::Geo(_)) => match geo_sorted {
+                Some(earlier_sort_index) => {
+                    canonicalization_actions.push(CanonicalizationAction::RemovedDuplicate {
+                        earlier_occurrence: RankingRuleSource::Sort {
+                            criterion_index,
+                            sort_index: earlier_sort_index,
+                        },
+                        removed_occurrence: source,
+                    })
+                }
+                None => {
+                    geo_sorted = Some(sort_index);
+                    canonical_sort.push(asc_desc);
+                }
+            },
+        }
+    }
+    canonical_sort
+}
+
+fn canonicalize_criterion(
+    criterion: Criterion,
+    criterion_index: usize,
+    terms_matching_strategy: TermsMatchingStrategy,
+    words: &mut Option<Option<usize>>,
+    canonicalization_actions: &mut Vec<CanonicalizationAction>,
+    canonical_criteria: &mut Vec<Criterion>,
+    rule: &mut Option<usize>,
+) {
+    *words = match (terms_matching_strategy, words.take()) {
+        (TermsMatchingStrategy::All, words) => words,
+        (_, None) => {
+            // inject words
+            canonicalization_actions.push(CanonicalizationAction::PrependedWords {
+                prepended_index: RankingRuleSource::Criterion(criterion_index),
+            });
+            canonical_criteria.push(Criterion::Words);
+            Some(None)
+        }
+        (_, words) => words,
+    };
+    if let Some(previous_index) = *rule {
+        canonicalization_actions.push(CanonicalizationAction::RemovedDuplicate {
+            earlier_occurrence: RankingRuleSource::Criterion(previous_index),
+            removed_occurrence: RankingRuleSource::Criterion(criterion_index),
+        });
+    } else {
+        *rule = Some(criterion_index);
+        canonical_criteria.push(criterion)
+    }
+}
+
+#[derive(Debug, Clone, Copy, PartialEq, Eq)]
+enum RankingRuleKind {
+    Relevancy,
+    AscendingSort,
+    DescendingSort,
+    AscendingGeoSort,
+    DescendingGeoSort,
+}
+
+#[derive(Debug, Clone, Copy)]
+pub struct RankingRule {
+    source: RankingRuleSource,
+    kind: RankingRuleKind,
+}
+
+#[derive(Debug, Clone, Copy)]
+pub enum RankingRuleSource {
+    Criterion(usize),
+    CoalescedCriteria(usize, usize),
+    Sort { criterion_index: usize, sort_index: usize },
+}
+
+impl RankingRuleSource {
+    fn rule_name(&self, criteria: &[Criterion], sort: &Option<Vec<AscDesc>>) -> String {
+        match self {
+            RankingRuleSource::Criterion(criterion_index) => criteria
+                .get(*criterion_index)
+                .map(|c| c.to_string())
+                .unwrap_or_else(|| "unknown".into()),
+            RankingRuleSource::CoalescedCriteria(begin, end) => {
+                let rules: Vec<_> = criteria
+                    .get(*begin..=*end)
+                    .iter()
+                    .flat_map(|c| c.iter())
+                    .map(|c| c.to_string())
+                    .collect();
+                rules.join(", ")
+            }
+            RankingRuleSource::Sort { criterion_index: _, sort_index } => {
+                match sort.as_deref().and_then(|sort| sort.get(*sort_index)) {
+                    Some(sort) => match sort {
+                        AscDesc::Asc(Member::Field(field_name)) => format!("{field_name}:asc"),
+                        AscDesc::Desc(Member::Field(field_name)) => {
+                            format!("{field_name}:desc")
+                        }
+                        AscDesc::Asc(Member::Geo(_)) => "_geo(..):asc".to_string(),
+                        AscDesc::Desc(Member::Geo(_)) => "_geo(..):desc".to_string(),
+                    },
+                    None => "unknown".into(),
+                }
+            }
+        }
+    }
+
+    fn rule_position(&self) -> String {
+        match self {
+            RankingRuleSource::Criterion(criterion_index) => {
+                format!("#{criterion_index} in ranking rules")
+            }
+            RankingRuleSource::CoalescedCriteria(begin, end) => {
+                format!("#{begin} to #{end} in ranking rules")
+            }
+            RankingRuleSource::Sort { criterion_index, sort_index } => format!(
+                "#{sort_index} in `query.sort` (as `sort` is #{criterion_index} in ranking rules)"
+            ),
+        }
+    }
+}
+
+impl RankingRule {
+    fn from_criterion<'a>(
+        criterion_index: usize,
+        criterion: &'a Criterion,
+        sort: &'a Option<Vec<AscDesc>>,
+    ) -> impl Iterator<Item = Self> + 'a {
+        let kind = match criterion {
+            Criterion::Words
+            | Criterion::Typo
+            | Criterion::Proximity
+            | Criterion::Attribute
+            | Criterion::Exactness => RankingRuleKind::Relevancy,
+            Criterion::Asc(s) if s == "_geo" => RankingRuleKind::AscendingGeoSort,
+
+            Criterion::Asc(_) => RankingRuleKind::AscendingSort,
+            Criterion::Desc(s) if s == "_geo" => RankingRuleKind::DescendingGeoSort,
+
+            Criterion::Desc(_) => RankingRuleKind::DescendingSort,
+            Criterion::Sort => {
+                return either::Right(sort.iter().flatten().enumerate().map(
+                    move |(rule_index, asc_desc)| {
+                        Self::from_asc_desc(asc_desc, criterion_index, rule_index)
+                    },
+                ))
+            }
+        };
+
+        either::Left(std::iter::once(Self {
+            source: RankingRuleSource::Criterion(criterion_index),
+            kind,
+        }))
+    }
+
+    fn from_asc_desc(asc_desc: &AscDesc, sort_index: usize, rule_index_in_sort: usize) -> Self {
+        let kind = match asc_desc {
+            AscDesc::Asc(Member::Field(_)) => RankingRuleKind::AscendingSort,
+            AscDesc::Desc(Member::Field(_)) => RankingRuleKind::DescendingSort,
+            AscDesc::Asc(Member::Geo(_)) => RankingRuleKind::AscendingGeoSort,
+            AscDesc::Desc(Member::Geo(_)) => RankingRuleKind::DescendingGeoSort,
+        };
+        Self {
+            source: RankingRuleSource::Sort {
+                criterion_index: sort_index,
+                sort_index: rule_index_in_sort,
+            },
+            kind,
+        }
+    }
+
+    fn as_string(
+        &self,
+        canonical_criteria: &[Criterion],
+        canonical_sort: &Option<Vec<AscDesc>>,
+        query_index: usize,
+        index_uid: &str,
+    ) -> String {
+        let kind = match self.kind {
+            RankingRuleKind::Relevancy => "relevancy",
+            RankingRuleKind::AscendingSort => "ascending sort",
+            RankingRuleKind::DescendingSort => "descending sort",
+            RankingRuleKind::AscendingGeoSort => "ascending geo sort",
+            RankingRuleKind::DescendingGeoSort => "descending geo sort",
+        };
+        let rules = self.fetch_from_source(canonical_criteria, canonical_sort);
+
+        let source = match self.source {
+            RankingRuleSource::Criterion(criterion_index) => format!("`queries[{query_index}]`, `{index_uid}.rankingRules[{criterion_index}]`"),
+            RankingRuleSource::CoalescedCriteria(begin, end) => format!("`queries[{query_index}]`, `{index_uid}.rankingRules[{begin}..={end}]`"),
+            RankingRuleSource::Sort { criterion_index, sort_index } => format!("`queries[{query_index}].sort[{sort_index}]`, `{index_uid}.rankingRules[{criterion_index}]`"),
+        };
+
+        format!("{source}: {kind} {rules}")
+    }
+
+    fn fetch_from_source(
+        &self,
+        canonical_criteria: &[Criterion],
+        canonical_sort: &Option<Vec<AscDesc>>,
+    ) -> String {
+        let rule_name = match self.source {
+            RankingRuleSource::Criterion(index) => {
+                canonical_criteria.get(index).map(|criterion| criterion.to_string())
+            }
+            RankingRuleSource::CoalescedCriteria(begin, end) => {
+                let rules: Vec<String> = canonical_criteria
+                    .get(begin..=end)
+                    .into_iter()
+                    .flat_map(|criteria| criteria.iter())
+                    .map(|criterion| criterion.to_string())
+                    .collect();
+
+                (!rules.is_empty()).then_some(rules.join(", "))
+            }
+            RankingRuleSource::Sort { criterion_index: _, sort_index } => canonical_sort
+                .as_deref()
+                .and_then(|canonical_sort| canonical_sort.get(sort_index))
+                .and_then(|asc_desc: &AscDesc| match asc_desc {
+                    AscDesc::Asc(Member::Field(s)) | AscDesc::Desc(Member::Field(s)) => {
+                        Some(format!("on field `{s}`"))
+                    }
+                    _ => None,
+                }),
+        };
+
+        let rule_name = rule_name.unwrap_or_else(|| "default".into());
+
+        format!("rule(s) {rule_name}")
+    }
+}

crates/meilisearch/src/search_queue.rs

@@ -0,0 +1,170 @@
+//! This file implements a queue of searches to process and the ability to control how many searches can be run in parallel.
+//! We need this because we don't want to process more search requests than we have cores.
+//! That slows down everything and consumes RAM for no reason.
+//! The steps to do a search are to get the `SearchQueue` data structure and try to get a search permit.
+//! This can fail if the queue is full, and we need to drop your search request to register a new one.
+//!
+//! ### How to do a search request
+//!
+//! In order to do a search request you should try to get a search permit.
+//! Retrieve the `SearchQueue` structure from actix-web (`search_queue: Data<SearchQueue>`)
+//! and right before processing the search, calls the `SearchQueue::try_get_search_permit` method: `search_queue.try_get_search_permit().await?;`
+//!
+//! What is going to happen at this point is that you're going to send a oneshot::Sender over an async mpsc channel.
+//! Then, the queue/scheduler is going to either:
+//! - Drop your oneshot channel => that means there are too many searches going on, and yours won't be executed.
+//!                                You should exit and free all the RAM you use ASAP.
+//! - Sends you a Permit => that will unlock the method, and you will be able to process your search.
+//!                         And should drop the Permit only once you have freed all the RAM consumed by the method.
+
+use std::num::NonZeroUsize;
+use std::time::Duration;
+
+use rand::rngs::StdRng;
+use rand::{Rng, SeedableRng};
+use tokio::sync::{mpsc, oneshot};
+
+use crate::error::MeilisearchHttpError;
+
+#[derive(Debug)]
+pub struct SearchQueue {
+    sender: mpsc::Sender<oneshot::Sender<Permit>>,
+    capacity: usize,
+    /// If we have waited longer than this to get a permit, we should abort the search request entirely.
+    /// The client probably already closed the connection, but we have no way to find out.
+    time_to_abort: Duration,
+}
+
+/// You should only run search requests while holding this permit.
+/// Once it's dropped, a new search request will be able to process.
+/// You should always try to drop the permit yourself calling the `drop` async method on it.
+#[derive(Debug)]
+pub struct Permit {
+    sender: mpsc::Sender<()>,
+}
+
+impl Permit {
+    /// Drop the permit giving back on permit to the search queue.
+    pub async fn drop(self) {
+        // if the channel is closed then the whole instance is down
+        let _ = self.sender.send(()).await;
+    }
+}
+
+impl Drop for Permit {
+    /// The implicit drop implementation can still be called in multiple cases:
+    /// - We forgot to call the explicit one somewhere => this should be fixed on our side asap
+    /// - The future is cancelled while running and the permit dropped with it
+    fn drop(&mut self) {
+        let sender = self.sender.clone();
+        // if the channel is closed then the whole instance is down
+        std::mem::drop(tokio::spawn(async move { sender.send(()).await }));
+    }
+}
+
+impl SearchQueue {
+    pub fn new(capacity: usize, paralellism: NonZeroUsize) -> Self {
+        // Search requests are going to wait until we're available anyway,
+        // so let's not allocate any RAM and keep a capacity of 1.
+        let (sender, receiver) = mpsc::channel(1);
+
+        tokio::task::spawn(Self::run(capacity, paralellism, receiver));
+        Self { sender, capacity, time_to_abort: Duration::from_secs(60) }
+    }
+
+    pub fn with_time_to_abort(self, time_to_abort: Duration) -> Self {
+        Self { time_to_abort, ..self }
+    }
+
+    /// This function is the main loop, it's in charge on scheduling which search request should execute first and
+    /// how many should executes at the same time.
+    ///
+    /// It **must never** panic or exit.
+    async fn run(
+        capacity: usize,
+        parallelism: NonZeroUsize,
+        mut receive_new_searches: mpsc::Receiver<oneshot::Sender<Permit>>,
+    ) {
+        let mut queue: Vec<oneshot::Sender<Permit>> = Default::default();
+        let mut rng: StdRng = StdRng::from_entropy();
+        let mut searches_running: usize = 0;
+        // By having a capacity of parallelism we ensures that every time a search finish it can release its RAM asap
+        let (sender, mut search_finished) = mpsc::channel(parallelism.into());
+
+        loop {
+            tokio::select! {
+                // biased select because we wants to free up space before trying to register new tasks
+                biased;
+                _ = search_finished.recv() => {
+                    searches_running = searches_running.saturating_sub(1);
+                    if !queue.is_empty() {
+                        // Can't panic: the queue wasn't empty thus the range isn't empty.
+                        let remove = rng.gen_range(0..queue.len());
+                        let channel = queue.swap_remove(remove);
+                        let _ = channel.send(Permit { sender: sender.clone() });
+                    }
+                },
+
+                search_request = receive_new_searches.recv() => {
+                    let search_request = match search_request {
+                        Some(search_request) => search_request,
+                        // This should never happen while actix-web is running, but it's not a reason to crash
+                        // and it can generate a lot of noise in the tests.
+                        None => continue,
+                    };
+
+                    if searches_running < usize::from(parallelism) && queue.is_empty() {
+                        searches_running += 1;
+                        // if the search requests die it's not a hard error on our side
+                        let _ = search_request.send(Permit { sender: sender.clone() });
+                        continue;
+                    } else if capacity == 0 {
+                        // in the very specific case where we have a capacity of zero
+                        // we must refuse the request straight away without going through
+                        // the queue stuff.
+                        drop(search_request);
+                        continue;
+
+                    } else if queue.len() >= capacity {
+                        let remove = rng.gen_range(0..queue.len());
+                        let thing = queue.swap_remove(remove); // this will drop the channel and notify the search that it won't be processed
+                        drop(thing);
+                    }
+                    queue.push(search_request);
+                },
+            }
+        }
+    }
+
+    /// Returns a search `Permit`.
+    /// It should be dropped as soon as you've freed all the RAM associated with the search request being processed.
+    pub async fn try_get_search_permit(&self) -> Result<Permit, MeilisearchHttpError> {
+        let now = std::time::Instant::now();
+        let (sender, receiver) = oneshot::channel();
+        self.sender.send(sender).await.map_err(|_| MeilisearchHttpError::SearchLimiterIsDown)?;
+        let permit = receiver
+            .await
+            .map_err(|_| MeilisearchHttpError::TooManySearchRequests(self.capacity))?;
+
+        // If we've been for more than one minute to get a search permit, it's better to simply
+        // abort the search request than spending time processing something were the client
+        // most certainly exited or got a timeout a long time ago.
+        // We may find a better solution in https://github.com/actix/actix-web/issues/3462.
+        if now.elapsed() > self.time_to_abort {
+            permit.drop().await;
+            Err(MeilisearchHttpError::TooManySearchRequests(self.capacity))
+        } else {
+            Ok(permit)
+        }
+    }
+
+    /// Returns `Ok(())` if everything seems normal.
+    /// Returns `Err(MeilisearchHttpError::SearchLimiterIsDown)` if the search limiter seems down.
+    pub fn health(&self) -> Result<(), MeilisearchHttpError> {
+        if self.sender.is_closed() {
+            Err(MeilisearchHttpError::SearchLimiterIsDown)
+        } else {
+            Ok(())
+        }
+    }
+}