ref(deletions): Use the same logic to schedule tasks (#103836)

This helps with code consistency.

Author: armenzg

src/sentry/api/helpers/group_index/delete.py

@@ -19,6 +19,7 @@
 from sentry.models.project import Project
 from sentry.signals import issue_deleted
 from sentry.utils.audit import create_audit_entry
+from sentry.utils.iterators import chunked
 
 from . import BULK_MUTATION_LIMIT, SearchFunction
 from .validators import ValidationError
@@ -85,14 +86,30 @@ def delete_group_list(
     GroupInbox.objects.filter(project_id=project.id, group_id__in=group_ids).delete()
 
     # Schedule a task per GROUP_CHUNK_SIZE batch of groups
-    for i in range(0, len(group_ids), GROUP_CHUNK_SIZE):
+    schedule_group_deletion_tasks(project.id, group_ids, transaction_id)
+
+
+def schedule_group_deletion_tasks(
+    project_id: int, group_ids: list[int], transaction_id: str | None = None
+) -> int:
+    """Schedule tasks to delete groups in batches of GROUP_CHUNK_SIZE.
+
+    :param project_id: The project ID.
+    :param group_ids: The list of group IDs to delete.
+    :param transaction_id: The transaction ID.
+    :return: The total number of tasks scheduled.
+    """
+    total_tasks = 0
+    for group_id_chunk in chunked(group_ids, GROUP_CHUNK_SIZE):
         delete_groups_for_project.apply_async(
             kwargs={
-                "project_id": project.id,
-                "object_ids": group_ids[i : i + GROUP_CHUNK_SIZE],
-                "transaction_id": str(transaction_id),
+                "project_id": project_id,
+                "object_ids": group_id_chunk,
+                "transaction_id": transaction_id or uuid4().hex,
             }
         )
+        total_tasks += 1
+    return total_tasks
 
 
 def create_audit_entries(

src/sentry/tasks/delete_pending_groups.py

@@ -1,12 +1,10 @@
 import logging
 from collections import defaultdict
 from datetime import timedelta
-from uuid import uuid4
 
 from django.utils import timezone
 
-from sentry.deletions.defaults.group import GROUP_CHUNK_SIZE
-from sentry.deletions.tasks.groups import delete_groups_for_project
+from sentry.api.helpers.group_index.delete import schedule_group_deletion_tasks
 from sentry.models.group import Group, GroupStatus
 from sentry.silo.base import SiloMode
 from sentry.tasks.base import instrumented_task
@@ -41,9 +39,7 @@ def delete_pending_groups() -> None:
     """
     statuses_to_delete = [GroupStatus.PENDING_DELETION, GroupStatus.DELETION_IN_PROGRESS]
 
-    # XXX: If needed add a partial index with the status and last_seen fields
-    # This can timeout for lack of an index on the status field
-    # Not using the last_seen index to avoid the lack of composite index on status and last_seen
+    # Just using status to take advantage of the status DB index
     groups = Group.objects.filter(status__in=statuses_to_delete).values_list(
         "id", "project_id", "last_seen"
     )[:BATCH_LIMIT]
@@ -71,19 +67,7 @@ def delete_pending_groups() -> None:
     )
 
     for project_id, group_ids in groups_by_project.items():
-        # Schedule deletion tasks in chunks of GROUP_CHUNK_SIZE
-        for i in range(0, len(group_ids), GROUP_CHUNK_SIZE):
-            chunk = group_ids[i : i + GROUP_CHUNK_SIZE]
-            transaction_id = str(uuid4())
-
-            delete_groups_for_project.apply_async(
-                kwargs={
-                    "project_id": project_id,
-                    "object_ids": chunk,
-                    "transaction_id": transaction_id,
-                }
-            )
-            total_tasks += 1
+        total_tasks += schedule_group_deletion_tasks(project_id, group_ids)
 
     metrics.incr("delete_pending_groups.groups_scheduled", amount=total_groups, sample_rate=1.0)
     metrics.incr("delete_pending_groups.tasks_scheduled", amount=total_tasks, sample_rate=1.0)

tests/sentry/tasks/test_delete_pending_groups.py

@@ -55,7 +55,7 @@ def test_schedules_only_groups_within_valid_date_range(self) -> None:
         )
 
         with patch(
-            "sentry.tasks.delete_pending_groups.delete_groups_for_project.apply_async"
+            "sentry.api.helpers.group_index.delete.delete_groups_for_project.apply_async"
         ) as mock_delete_task:
             delete_pending_groups()
 
@@ -76,7 +76,7 @@ def test_schedules_only_groups_within_valid_date_range(self) -> None:
             wrong_status.id,
         ]
 
-    @patch("sentry.tasks.delete_pending_groups.delete_groups_for_project.apply_async")
+    @patch("sentry.api.helpers.group_index.delete.delete_groups_for_project.apply_async")
     def test_groups_by_project(self, mock_delete_task: MagicMock) -> None:
         """Test that groups are properly grouped by project when scheduling deletion."""
         project1 = self.create_project()
@@ -109,8 +109,8 @@ def test_groups_by_project(self, mock_delete_task: MagicMock) -> None:
             elif call_kwargs["project_id"] == project2.id:
                 assert set(call_kwargs["object_ids"]) == {group3.id}
 
-    @patch("sentry.tasks.delete_pending_groups.GROUP_CHUNK_SIZE", new=10)
-    @patch("sentry.tasks.delete_pending_groups.delete_groups_for_project.apply_async")
+    @patch("sentry.api.helpers.group_index.delete.GROUP_CHUNK_SIZE", 10)
+    @patch("sentry.api.helpers.group_index.delete.delete_groups_for_project.apply_async")
     @patch("sentry.tasks.delete_pending_groups.metrics.incr")
     def test_chunks_large_batches(
         self,