Skip to content

:hedgehog: PostHog is the leading platform for building self-driving products. Our developer tools – AI observability, analytics, session replay, flags, experiments, error tracking, logs, and more – capture all the context agents need to diagnose problems, uncover opportunities, and ship fixes. Steer it all from Slack, web, desktop, or the MCP.

perf(replay-vision): count highlight group reach from member rows (#125500)

Tue Haulund0ad9cb2d3bec
diff --git a/products/replay_vision/backend/tests/test_api.py b/products/replay_vision/backend/tests/test_api.py
index 4b2927a22720..80c301b40f7d 100644
--- a/products/replay_vision/backend/tests/test_api.py
+++ b/products/replay_vision/backend/tests/test_api.py
@@ -4957,8 +4957,11 @@ class TestHighlightFeedAPI(_VisionAPITestCase):
self._ranked_observation(scanner, "copy", 400, "failure", -0.5, group=str(oldest.id))
for index in range(10):
self._ranked_observation(scanner, f"routine-{index}", index)
+ # A row outside the reader's scanner selection never counts, even under the same group id.
+ other = self._create_scanner(name="other")
+ self._ranked_observation(other, "elsewhere", 300, "failure", 0.9, group=str(oldest.id))
- items = self.client.get(self.feed_url).json()["results"]
+ items = self.client.get(self.feed_url, {"scanner_ids": str(scanner.id)}).json()["results"]
picks = [item for item in items if item["reason"]["kind"] == "embedding_pick"]
self.assertEqual([item["observation"]["session_id"] for item in picks], ["oldest-failure"])
self.assertEqual(picks[0]["reason"]["similar_count"], 1)
diff --git a/products/replay_vision/backend/watch_rank.py b/products/replay_vision/backend/watch_rank.py
index 9e8abdce741c..cb2b5df02879 100644
--- a/products/replay_vision/backend/watch_rank.py
+++ b/products/replay_vision/backend/watch_rank.py
@@ -396,16 +396,25 @@ def select_embedding_picks(
.filter(member_root__in=list(roots))
.values_list("id", flat=True)
)
+ # Read by id with the rank checks in Python, since the rank filters steer Postgres onto the partial rank indexes.
+ readable_scanners = {str(scanner_id) for scanner_id in scanner_ids}
+ members = (
+ accessible_observations(access, team_id, walkable.filter(id__in=member_ids))
+ .values_list("id", "scanner_id", "watch_rank", person)
+ .iterator()
+ )
+ reach_sessions: Counter[str] = Counter()
+ reach_people: dict[str, set[str]] = {}
+ for member_id, scanner_id, watch_rank, member_person in members:
+ rank = current_rank({"watch_rank": watch_rank})
+ if str(scanner_id) not in readable_scanners or rank is None or rank.get("category") not in categories:
+ continue
+ group = rank.get("group")
+ root = str(group) if group is not None else str(member_id)
+ reach_sessions[root] += 1
+ reach_people.setdefault(root, set()).add(member_person)
group_reach = {
- row["watch_group_root"]: row
- for row in accessible_observations(
- access,
- team_id,
- current_ranks.filter(id__in=member_ids, watch_category__in=list(categories)),
- )
- .annotate(watch_group_root=group_root)
- .values("watch_group_root")
- .annotate(sessions=Count("id"), people=Count(person, distinct=True))
+ root: {"sessions": sessions, "people": len(reach_people[root])} for root, sessions in reach_sessions.items()
}
# The row-gate reads each row's snapshot JSON, so it runs on the picks only.
rows = cast(