server: call the webhook asynchronously in event triggers

CHANGELOG.md

@@ -6,6 +6,7 @@
 
 (Add entries here in the order of: server, console, cli, docs, others)
 
+- server: process events generated by the event triggers asynchronously (close #5189) (#5352)
 - console: display line number that error originated from in GraphQL editor (close #4849) (#4942)
 
 ## `v1.3.0-beta.4`

server/src-lib/Hasura/Eventing/EventTrigger.hs

@@ -41,7 +41,7 @@ module Hasura.Eventing.EventTrigger
   ) where
 
 
-import           Control.Concurrent.Async    (wait, withAsync)
+import           Control.Concurrent.Async    (async, link, wait, withAsync)
 import           Control.Concurrent.Extended (sleep)
 import           Control.Concurrent.STM.TVar
 import           Control.Monad.Catch         (MonadMask, bracket_)
@@ -189,10 +189,13 @@ processEventQueue logger logenv httpMgr pool getSchemaCache eeCtx@EventEngineCtx
       eventsNext <- withAsync popEventsBatch $ \eventsNextA -> do
         -- process approximately in order, minding HASURA_GRAPHQL_EVENTS_HTTP_POOL_SIZE:
         forM_ events $ \event -> do
-             runReaderT (withEventEngineCtx eeCtx $ (processEvent event)) (logger, httpMgr)
+             t <- async $ runReaderT (withEventEngineCtx eeCtx $ (processEvent event)) (logger, httpMgr)
              -- removing an event from the _eeCtxLockedEvents after the event has
              -- been processed
              removeEventFromLockedEvents (eId event) leEvents
+             link t
+
+        -- return when next batch ready; some 'processEvent' threads may be running.
         wait eventsNextA
 
       let lenEvents = length events

server/tests-py/queries/event_triggers/async_execution/setup.yaml

@@ -0,0 +1,26 @@
+type: bulk
+args:
+- type: run_sql
+  args:
+    sql: |
+      create table hge_tests.test_t1(
+          c1 int,
+          c2 text
+      );
+- type: track_table
+  args:
+    schema: hge_tests
+    name: test_t1
+- type: create_event_trigger
+  args:
+    name: t1_timeout_long
+    table:
+      schema: hge_tests
+      name: test_t1
+    insert:
+      columns: "*"
+    update:
+      columns: "*"
+    delete:
+      columns: "*"
+    webhook: http://127.0.0.1:5592/timeout_long

server/tests-py/queries/event_triggers/async_execution/teardown.yaml

@@ -0,0 +1,9 @@
+type: bulk
+args:
+- type: delete_event_trigger
+  args:
+    name: t1_timeout_long
+- type: run_sql
+  args:
+    sql: |
+      drop table hge_tests.test_t1;

server/tests-py/test_events.py

@@ -588,3 +588,37 @@ def test_basic(self, hge_ctx, evts_webhook):
         assert st_code == 200, resp
         st_code, resp = hge_ctx.v1q_f('queries/event_triggers/manual_events/disabled.yaml')
         assert st_code == 400, resp
+
+@usefixtures('per_method_tests_db_state')
+class TestEventsAsynchronousExecution(object):
+
+    @classmethod
+    def dir(cls):
+        return 'queries/event_triggers/async_execution'
+
+    def test_async_execution(self,hge_ctx,evts_webhook):
+        """
+        A test to check if the events generated by the graphql-engine are
+        processed asynchronously. This test measures the time taken to process
+        all the events and that time should definitely be lesser than the time
+        taken if the events were to be executed sequentially.
+
+        This test inserts 5 rows and the webhook(/timeout_long) takes
+        ~5 seconds to process one request. So, if the graphql-engine
+        were to process the events sequentially it will take 5 * 5 = 25 seconds.
+        Theorotically, all the events should have been processed in ~5 seconds,
+        adding a 5 seconds buffer to the comparision, so that this test
+        doesn't flake in the CI.
+        """
+        table = {"schema": "hge_tests", "name": "test_t1"}
+
+        payload = range(1,6)
+        rows = list(map(lambda x: {"c1": x, "c2": "hello"}, payload))
+        st_code, resp = insert_many(hge_ctx, table, rows)
+        start_time = time.perf_counter()
+        assert st_code == 200, resp
+        for i in range(1,6):
+            _ = evts_webhook.get_event(7) # webhook takes 5 seconds to process a request
+        end_time = time.perf_counter()
+        time_elapsed = end_time - start_time
+        assert time_elapsed < 10