perf(alerts): Limit GroupEvent processing to relevant groups

Author: kcons

src/sentry/rules/processing/delayed_processing.py

@@ -297,12 +297,17 @@ def get_group_to_groupevent(
     groups = Group.objects.filter(id__in=group_ids)
     group_id_to_group = {group.id: group for group in groups}
 
+    # Filter down to only the event data for the groups we've been asked to process.
+    relevant_rulegroup_to_event_data = {
+        key: value for key, value in parsed_rulegroup_to_event_data.items() if key[1] in group_ids
+    }
+
     # Use a list comprehension for event_ids
     event_ids: list[str] = [
         event_id
         for event_id in (
             instance_data.get("event_id")
-            for instance_data in parsed_rulegroup_to_event_data.values()
+            for instance_data in relevant_rulegroup_to_event_data.values()
         )
         if event_id is not None
     ]
@@ -312,7 +317,7 @@ def get_group_to_groupevent(
         occurrence_id
         for occurrence_id in (
             instance_data.get("occurrence_id")
-            for instance_data in parsed_rulegroup_to_event_data.values()
+            for instance_data in relevant_rulegroup_to_event_data.values()
         )
         if occurrence_id is not None
     ]
@@ -325,7 +330,7 @@ def get_group_to_groupevent(
     }
 
     return build_group_to_groupevent(
-        parsed_rulegroup_to_event_data,
+        relevant_rulegroup_to_event_data,
         bulk_event_id_to_events,
         bulk_occurrence_id_to_occurrence,
         group_id_to_group,

tests/sentry/rules/processing/test_delayed_processing.py

@@ -378,6 +378,30 @@ def test_invalid_group_ids(self):
         result = get_group_to_groupevent(self.parsed_data, self.project.id, {0})
         assert len(result) == 0
 
+    def test_filtered_group_ids(self):
+        """Test that get_group_to_groupevent only requests events for specified group IDs."""
+        # Track which event IDs are requested
+        requested_event_ids = set()
+
+        original_bulk_fetch_events = bulk_fetch_events
+
+        def mock_bulk_fetch_events(event_ids, project_id):
+            requested_event_ids.update(event_ids)
+            return original_bulk_fetch_events(event_ids, project_id)
+
+        with patch(
+            "sentry.rules.processing.delayed_processing.bulk_fetch_events",
+            side_effect=mock_bulk_fetch_events,
+        ):
+            # Call get_group_to_groupevent with only group1
+            result = get_group_to_groupevent(self.parsed_data, self.project.id, {self.group1.id})
+
+            # Verify only event1 was requested
+            assert requested_event_ids == {self.event1.event_id}
+            assert len(result) == 1
+            assert result[self.group1] == self.event1
+            assert self.group2 not in result
+
 
 class GetRulesToFireTest(TestCase):
     def setUp(self):