feat(issue-search): implements x-hits and cursor logic for snuba only search (#70089)

This PR makes sure we properly set the `X-Hits` and pagination logic for
the new snuba only search. One thing I want to call out is that there
might be issues when issues are pending deletion or being merged. We may
return less than a full page of results in that case. This problem is
likely worse than the postgres only search because we aren't checking PG
for issue state before hitting Snuba.

Note that I am not sure about some of the logic overriding the default
behavior for `has_results`. The existing logic is somewhat convoluted
and I'm not sure we actually need it.

I also added more filters that we currently don't support for snuba only
search. Some of them like `firstSeen` should be easy to support while
others are harder.

src/sentry/issues/endpoints/organization_group_index.py

@@ -56,6 +56,12 @@ UNSUPPORTED_SNUBA_FILTERS = [
     "subscribed_by",
     "regressed_in_release",
     "issue.priority",
+    "firstRelease",
+    "firstSeen",
+    "lastSeen",
+    "release.build",
+    "release.package",
+    "release.dist",
 ]
 
 

src/sentry/search/snuba/executors.py

@@ -1618,7 +1618,7 @@ class GroupAttributesPostgresSnubaQueryExecutor(PostgresSnubaQueryExecutor):
         )
 
         data = []
-        count = 0
+        count: int = 0
         # get the query data and the query counts
         k = 0
         for _ in range(len(entities_to_check)):
@@ -1628,20 +1628,24 @@ class GroupAttributesPostgresSnubaQueryExecutor(PostgresSnubaQueryExecutor):
                 count += bulk_result[k]["data"][0]["count"]
             k += 1
 
-        hits = 0
         paginator_results = SequencePaginator(
             [(row[self.sort_strategies[sort_by]], row["g.group_id"]) for row in data],
             reverse=True,
             **paginator_options,
-        ).get_result(limit, cursor, known_hits=hits, max_hits=max_hits)
-
-        # We filter against `group_queryset` here so that we recheck all conditions in Postgres.
-        # Since replay between Postgres and Clickhouse can happen, we might get back results that
-        # have changed state in Postgres. By rechecking them we guarantee than any returned results
-        # have the correct state.
-        # TODO: This can result in us returning less than a full page of results, but shouldn't
-        # affect cursors. If we want to, we can iterate and query snuba until we manage to get a
-        # full page. In practice, this will likely only skip a couple of results at worst, and
+        ).get_result(limit, cursor, known_hits=count, max_hits=max_hits)
+
+        # TODO: do we need to set has_results for the next cursor?
+
+        if cursor is not None and (not cursor.is_prev or len(paginator_results.results) > 0):
+            # If the user passed a cursor, and it isn't already a 0 result `is_prev`
+            # cursor, then it's worth allowing them to go back a page to check for
+            # more results.
+            paginator_results.prev.has_results = True
+
+        # We filter against `group_queryset` here to ensure we only return groups that aren't being migrated
+        # or being deleted before snuba is updated. This can result in us returning less than a full page of
+        # results, but we can't do much about that without a more complex solution such as asking for more results
+        # In practice, this will likely only skip a couple of results at worst, and
         # probably not be noticeable to the user, so holding off for now to reduce complexity.
 
         groups = group_queryset.in_bulk(paginator_results.results)

tests/sentry/issues/endpoints/test_organization_group_index.py

@@ -2925,6 +2925,60 @@ class GroupListTest(APITestCase, SnubaTestCase, SearchIssueTestMixin):
         )
         assert len(response.data) == 0
 
+    @override_options({"issues.group_attributes.send_kafka": True})
+    def test_pagination_and_x_hits_header(self):
+        # Create 30 issues
+        for i in range(30):
+            self.store_event(
+                data={
+                    "timestamp": iso_format(before_now(seconds=i)),
+                    "fingerprint": [f"group-{i}"],
+                },
+                project_id=self.project.id,
+            )
+
+        self.login_as(user=self.user)
+
+        # Request the first page with a limit of 10
+        response = self.get_success_response(limit=10, useGroupSnubaDataset=1)
+        assert response.status_code == 200
+        assert len(response.data) == 10
+        assert response.headers.get("X-Hits") == "30"
+        assert "Link" in response.headers
+
+        # Parse the Link header to get the cursor for the next page
+        header_links = parse_link_header(response.headers["Link"])
+        next_obj = [link for link in header_links.values() if link["rel"] == "next"][0]
+        assert next_obj["results"] == "true"
+        cursor = next_obj["cursor"]
+        prev_obj = [link for link in header_links.values() if link["rel"] == "previous"][0]
+        assert prev_obj["results"] == "false"
+
+        # Request the second page using the cursor
+        response = self.get_success_response(limit=10, cursor=cursor, useGroupSnubaDataset=1)
+        assert response.status_code == 200
+        assert len(response.data) == 10
+
+        # Check for the presence of the next cursor
+        header_links = parse_link_header(response.headers["Link"])
+        next_obj = [link for link in header_links.values() if link["rel"] == "next"][0]
+        assert next_obj["results"] == "true"
+        cursor = next_obj["cursor"]
+        prev_obj = [link for link in header_links.values() if link["rel"] == "previous"][0]
+        assert prev_obj["results"] == "true"
+
+        # Request the third page using the cursor
+        response = self.get_success_response(limit=10, cursor=cursor, useGroupSnubaDataset=1)
+        assert response.status_code == 200
+        assert len(response.data) == 10
+
+        # Check that there is no next page
+        header_links = parse_link_header(response.headers["Link"])
+        next_obj = [link for link in header_links.values() if link["rel"] == "next"][0]
+        assert next_obj["results"] == "false"
+        prev_obj = [link for link in header_links.values() if link["rel"] == "previous"][0]
+        assert prev_obj["results"] == "true"
+
 
 class GroupUpdateTest(APITestCase, SnubaTestCase):
     endpoint = "sentry-api-0-organization-group-index"