feat(frontend): add max_batch_queries_per_frontend_node config (#15574

)

risedev.yml

@@ -925,6 +925,18 @@ profile:
       - use: prometheus
       - use: grafana
 
+  full-with-batch-query-limit:
+    config-path: src/config/full-with-batch-query-limit.toml
+    steps:
+        - use: minio
+        - use: etcd
+        - use: meta-node
+        - use: compute-node
+        - use: frontend
+        - use: compactor
+        - use: prometheus
+        - use: grafana
+
 compose:
   risingwave: "ghcr.io/risingwavelabs/risingwave:latest"
   prometheus: "prom/prometheus:latest"

src/common/src/config.rs

@@ -478,9 +478,14 @@ pub struct BatchConfig {
     #[config_doc(omitted)]
     pub developer: BatchDeveloperConfig,
 
+    /// This is the max number of queries per sql session.
     #[serde(default)]
     pub distributed_query_limit: Option<u64>,
 
+    /// This is the max number of batch queries per frontend node.
+    #[serde(default)]
+    pub max_batch_queries_per_frontend_node: Option<u64>,
+
     #[serde(default = "default::batch::enable_barrier_read")]
     pub enable_barrier_read: bool,
 

src/config/docs.md

@@ -6,9 +6,10 @@ This page is automatically generated by `./risedev generate-example-config`
 
 | Config | Description | Default |
 |--------|-------------|---------|
-| distributed_query_limit |  |  |
+| distributed_query_limit |  This is the max number of queries per sql session. |  |
 | enable_barrier_read |  | false |
 | frontend_compute_runtime_worker_threads |  frontend compute runtime worker threads | 4 |
+| max_batch_queries_per_frontend_node |  This is the max number of batch queries per frontend node. |  |
 | statement_timeout_in_sec |  Timeout for a batch query in seconds. | 3600 |
 | worker_threads_num |  The thread number of the batch task runtime in the compute node. The default value is  decided by `tokio`. |  |
 

src/config/full-with-batch-query-limit.toml

@@ -0,0 +1,2 @@
+[batch]
+max_batch_queries_per_frontend_node = 2

src/frontend/src/scheduler/distributed/query.rs

@@ -73,6 +73,8 @@ pub struct QueryExecution {
     shutdown_tx: Sender<QueryMessage>,
     /// Identified by process_id, secret_key. Query in the same session should have same key.
     pub session_id: SessionId,
+    /// Permit to execute the query. Once query finishes execution, this is dropped.
+    pub permit: Option<tokio::sync::OwnedSemaphorePermit>,
 }
 
 struct QueryRunner {
@@ -94,7 +96,11 @@ struct QueryRunner {
 
 impl QueryExecution {
     #[allow(clippy::too_many_arguments)]
-    pub fn new(query: Query, session_id: SessionId) -> Self {
+    pub fn new(
+        query: Query,
+        session_id: SessionId,
+        permit: Option<tokio::sync::OwnedSemaphorePermit>,
+    ) -> Self {
         let query = Arc::new(query);
         let (sender, receiver) = channel(100);
         let state = QueryState::Pending {
@@ -106,6 +112,7 @@ impl QueryExecution {
             state: RwLock::new(state),
             shutdown_tx: sender,
             session_id,
+            permit,
         }
     }
 
@@ -507,7 +514,7 @@ pub(crate) mod tests {
         let query = create_query().await;
         let query_id = query.query_id().clone();
         let pinned_snapshot = hummock_snapshot_manager.acquire();
-        let query_execution = Arc::new(QueryExecution::new(query, (0, 0)));
+        let query_execution = Arc::new(QueryExecution::new(query, (0, 0), None));
         let query_execution_info = Arc::new(RwLock::new(QueryExecutionInfo::new_from_map(
             HashMap::from([(query_id, query_execution.clone())]),
         )));

src/frontend/src/scheduler/distributed/query_manager.rs

@@ -25,6 +25,7 @@ use risingwave_common::session_config::QueryMode;
 use risingwave_pb::batch_plan::TaskOutputId;
 use risingwave_pb::common::HostAddress;
 use risingwave_rpc_client::ComputeClientPoolRef;
+use tokio::sync::OwnedSemaphorePermit;
 
 use super::stats::DistributedQueryMetrics;
 use super::QueryExecution;
@@ -131,7 +132,12 @@ pub struct QueryManager {
     catalog_reader: CatalogReader,
     query_execution_info: QueryExecutionInfoRef,
     pub query_metrics: Arc<DistributedQueryMetrics>,
+    /// Limit per session.
     disrtibuted_query_limit: Option<u64>,
+    /// Limits the number of concurrent distributed queries.
+    distributed_query_semaphore: Option<Arc<tokio::sync::Semaphore>>,
+    /// Total permitted distributed query number.
+    pub total_distributed_query_limit: Option<u64>,
 }
 
 impl QueryManager {
@@ -141,14 +147,39 @@ impl QueryManager {
         catalog_reader: CatalogReader,
         query_metrics: Arc<DistributedQueryMetrics>,
         disrtibuted_query_limit: Option<u64>,
+        total_distributed_query_limit: Option<u64>,
     ) -> Self {
+        let distributed_query_semaphore = total_distributed_query_limit
+            .map(|limit| Arc::new(tokio::sync::Semaphore::new(limit as usize)));
         Self {
             worker_node_manager,
             compute_client_pool,
             catalog_reader,
             query_execution_info: Arc::new(RwLock::new(QueryExecutionInfo::default())),
             query_metrics,
             disrtibuted_query_limit,
+            distributed_query_semaphore,
+            total_distributed_query_limit,
+        }
+    }
+
+    async fn get_permit(&self) -> SchedulerResult<Option<OwnedSemaphorePermit>> {
+        match self.distributed_query_semaphore {
+            Some(ref semaphore) => {
+                let permit = semaphore.clone().acquire_owned().await;
+                match permit {
+                    Ok(permit) => Ok(Some(permit)),
+                    Err(_) => {
+                        self.query_metrics.rejected_query_counter.inc();
+                        Err(crate::scheduler::SchedulerError::QueryReachLimit(
+                            QueryMode::Distributed,
+                            self.total_distributed_query_limit
+                                .expect("should have distributed query limit"),
+                        ))
+                    }
+                }
+            }
+            None => Ok(None),
         }
     }
 
@@ -167,7 +198,8 @@ impl QueryManager {
             ));
         }
         let query_id = query.query_id.clone();
-        let query_execution = Arc::new(QueryExecution::new(query, context.session().id()));
+        let permit = self.get_permit().await?;
+        let query_execution = Arc::new(QueryExecution::new(query, context.session().id(), permit));
 
         // Add queries status when begin.
         context

src/frontend/src/session.rs

@@ -173,6 +173,7 @@ impl FrontendEnv {
             catalog_reader.clone(),
             Arc::new(DistributedQueryMetrics::for_test()),
             None,
+            None,
         );
         let server_addr = HostAddr::try_from("127.0.0.1:4565").unwrap();
         let client_pool = Arc::new(ComputeClientPool::default());
@@ -277,6 +278,7 @@ impl FrontendEnv {
             catalog_reader.clone(),
             Arc::new(GLOBAL_DISTRIBUTED_QUERY_METRICS.clone()),
             batch_config.distributed_query_limit,
+            batch_config.max_batch_queries_per_frontend_node,
         );
 
         let user_info_manager = Arc::new(RwLock::new(UserInfoManager::default()));