[jzhu13]

Warning

Hold the merge. It moves new insight alerts onto the alerts platform, so it waits for the customer email about check-timing changes, and for #127773 (the logs version) to merge.

Problem

• A team with no insight alerts, including every new team, creates its first alert on the insight stack. It stays there until someone cuts the team over.
• The logs version of this fix is feat(logs): start a team's first logs alert on the alerts platform #127773. Its follow-ups ask for the same change on insight once the insight cutover begins.

Changes

• A new insight alert now starts on the alerts platform unless its team has other insight alerts and none of them is on the platform.
• A team with no other alert, and every new team, starts on the platform.
• A team on the platform keeps getting new alerts there, as before.
• A team whose alerts all stay on insight, such as one rolled back, keeps new alerts on insight.
• Alerts the platform cannot run (real-time, investigation agent, child environments, no matching delivery path) still stay on insight.
• Mechanical: has_begun_cutover becomes takes_new_alert, and tests that assumed a new team's alert stays on insight now create an insight-owned alert first.

How did you test this code?

Test rationale: extended test_a_new_alert_follows_its_team_and_starts_on_the_platform_in_a_team_without_alerts (renamed from test_a_new_alert_is_handed_over_only_in_a_team_on_the_platform) with a team that has no other alert. The old rule left that alert on insight.

• Ran products/alerts/backend/, posthog/tasks/alerts/test/, the alert interval API tests, products/metrics/backend/tests/test_alerts.py, and repo-wide mypy locally.
• Not checked: the dashboard and insight API tests that create alerts; CI covers them.

Release status

• No feature flag controls this change

Docs update

• None.

🤖 Agent context

Autonomy: Human-driven (agent-assisted)

Agent: Claude Code, Opus 5.5 (claude-opus-5-5)

Skills: /writing-tests, /writing-code-comments, /writing-pr-descriptions

• No duplicate: the open PR search found only feat(logs): start a team's first logs alert on the alerts platform #127773, the logs version, which this one copies.

More context for AI agents

• Decisions: the rule copies takes_new_alert from feat(logs): start a team's first logs alert on the alerts platform #127773 unchanged, so both sources answer "who owns a new alert" the same way.
• Follow-ups: the shadow backfill gap for new teams' first alerts closes once this merges, because those alerts start on the platform.
• Read closely: products/alerts/backend/platform_cutover.py takes_new_alert.

🤖 Generated with Claude Code

[trunk-io]

Merging to master in this repository is managed by Trunk.

• To merge this pull request, check the box to the left or comment /trunk merge below.

After your PR is submitted to the merge queue, this comment will be automatically updated with its status. If the PR fails, failure details will also be posted here

[github-actions]

🤖 CI report

⚠️ Trunk lane — backend Python lane

This PR is assigned to the backend Python lane. It runs backend Python tests and may merge in parallel with PRs in other lanes.

✅ Duplication (Python) — clean

New Python code duplication introduced by this branch. Fails at 70+ tokens in app code, or 150+ tokens when both copies live in test files. Advisory while the gate proves itself: extract a shared helper instead of copying.

✅ Duplication (TypeScript) — clean

New TypeScript code duplication introduced by this branch. Fails at 70+ tokens in app code, or 150+ tokens when both copies live in test files. Advisory while the gate proves itself: extract a shared helper instead of copying.

⚠️ Playwright — 1 flaky

🎭 Playwright report · View test results →

⚠️ 1 flaky test:

• Inline editing insight title via compact card popover (chromium)

These issues are not necessarily caused by your changes.
Annoyed by this section? Help fix flakies and failures and it will go green!

[parameterai]

Risk: No findings

This increment backdates a newly created platform-owned insight alert's shadow start (owner_changed_at) to the team's earliest existing platform cutover time, so the insight stack stops its redundant comparison checks for that alert when the team's week ends instead of starting a new week per alert. The platform copy still takes the real handover time, and delivery was already fully platform-owned, so the change only shortens a telemetry-only shadow window; I found no security impact in the new team_cutover_at query, the shadow_from plumbing, or their interaction with the shadow/rollback paths.

_{Sentinel reviewed 40c1776 · Review settings}

[stamphog]

Not approved — this change needs a human reviewer.

Re-add the stamphog label to request another review once you have addressed this.

The risky creation-path change has owning-team authorship, and no substantive review concerns are present; the trusted context also reports a clean current-head agent assessment. The explicit release hold remains outstanding: ask @jzhu13 to confirm its prerequisites and lift it before re-requesting.

• Author wrote 67% of the modified lines and has 24 merged PRs in these paths (familiarity MODERATE).
• PostHog Review (Standard) reviewed the current head: no findings.
• Outstanding maintainer hold: @jzhu13, an owning-team member, explicitly marks this PR as waiting for customer communication and PR #127773; no withdrawal is present. Have @jzhu13 confirm those prerequisites are complete and explicitly lift the hold before re-requesting.

Gate mechanics and policy version

Gate		Result
prerequisites	✓	all clear
deny-list	✓	no deny categories matched
size	✓	26L, 2F substantive, 86L/10F incl. docs/generated/snapshots — within ceiling
tier	✓	T1-agent / T1c-medium (86L, 10F, two-areas, feat)
stamphog 2.7.0		.stamphog/policy.yml @ 5c12e51 · reviewed head 5c12e51

[veria-ai]

Reviews

• 40c1776 — No new issues
• 52b57ec — No new issues
• 5c12e51 — No new issues

[posthog]

🤖 Agent review by Daniel's review agent - this is not Daniel's review.
Daniel has not read this PR yet. Treat these comments as automated findings, not as his opinion.

Result: no findings
Skills used: rs-review-pr (teammate mode), rs-adversarial-review, review-hog-perspective-logic-correctness, review-hog-perspective-performance-reliability
Head reviewed: 40c1776