fix(deletions): Switch to bulk deletions for GroupHistory

[armenzg]

This is how it used to be before #76608. Now that we have fixed the circular relationship (#102470) I believe we should give this approach another attempt.

Fixes SENTRY-5BX2:

For posterity:

OperationalError
canceling statement due to user request

SQL: SELECT "sentry_grouphistory"."id" AS "id" FROM "sentry_grouphistory" WHERE "sentry_grouphistory"."group_id" = %s ORDER BY 1 ASC LIMIT 100

The group in question has only been seen 4 times yet it has 905,729 rows of group history. This is caused by the priority of the group changing back and forth between high and medium priority. We're investigating the root cause of it.

[armenzg]

If it works, we can remove the custom class.

[codecov]

Codecov Report

❌ Patch coverage is 80.00000% with 1 line in your changes missing coverage. Please review.
✅ All tests successful. No failed tests found.

Files with missing lines	Patch %	Lines
src/sentry/deletions/__init__.py	75.00%	1 Missing ⚠️

Additional details and impacted files

@@             Coverage Diff             @@
##           master   #102580      +/-   ##
===========================================
+ Coverage   80.88%    80.89%   +0.01%     
===========================================
  Files        8861      8865       +4     
  Lines      390436    390860     +424     
  Branches    24832     24832              
===========================================
+ Hits       315786    316195     +409     
- Misses      74285     74300      +15     
  Partials      365       365              

[markstory]

Worth a shot now that the recursion issue is addressed.

[wedamija]

If the group has this many rows, there's a good chance that the query you have will still time out, since there is no index on group_id, id. I don't mind if you want to try, but I think the solution here would be to add that index if this doesn't work. You could even just do it temporarily until the deletions finish.

[wedamija]

If the group has this many rows, there's a good chance that the query you have will still time out, since there is no index on group_id, id. I don't mind if you want to try, but I think the solution here would be to add that index if this doesn't work. You could even just do it temporarily until the deletions finish.

Actually, how about this - just remove the ordering from your query? You don't care about the order here, you just want to delete rows.

chunk_ids = list(queryset.order_by("id").values_list("id", flat=True)[:100])

to

chunk_ids = list(queryset.values_list("id", flat=True)[:100])

[wedamija]

If the group has this many rows, there's a good chance that the query you have will still time out, since there is no index on group_id, id. I don't mind if you want to try, but I think the solution here would be to add that index if this doesn't work. You could even just do it temporarily until the deletions finish.

Actually, how about this - just remove the ordering from your query? You don't care about the order here, you just want to delete rows.

chunk_ids = list(queryset.order_by("id").values_list("id", flat=True)[:100])

to

chunk_ids = list(queryset.values_list("id", flat=True)[:100])

Sorry, I was looking at the sentry error that you linked here and this feedback was to fix the error seen there. Afaict, the BulkModelDeletionTask already does this, so lgtm

[armenzg]

Related code:

sentry/src/sentry/utils/query.py

Lines 272 to 294 in 2eecdd2

	def bulk_delete_objects(
	model: type[Model],
	limit: int = 10000,
	transaction_id: str | None = None,
	logger: logging.Logger | None = None,
	**filters: Any,
	) -> bool:
	qs = model.objects.filter(**filters).values_list("id")[:limit]
	delete_query = DeleteQuery(model)
	delete_query.add_q(Q(id__in=qs))
	n = delete_query.get_compiler(router.db_for_write(model)).execute_sql(ROW_COUNT)
	# `n` will be `None` if `qs` is an empty queryset
	has_more = n is not None and n > 0
	if has_more and logger is not None and _leaf_re.search(model.__name__) is None:
	logger.info(
	"object.delete.bulk_executed",
	extra=dict(filters, model=model.__name__, transaction_id=transaction_id),
	)
	return has_more