[developerz-ai]

What

health.md's release-run signal filtered gh run list by --branch "$t",
but release.yml:81-92 refuses any ref that is not refs/tags/v* and
release-tag.yml:193 dispatches every release at the tag — so no release
run carries a branch ref and a healthy publish read as zero runs, which
the row then called unhealthy.

The row now lists the last run and matches it to $t by headSha (or the
refs/tags/$t ref), and an empty list is a "no run yet", not a verdict.

Also keeps the applied SQL_ENQUEUE warmup in the jobs cross-process wake
test, which primes the statement shape before the SOON_MS measurement.

Acceptance

• .dz/briefs/health.md line 13 contains no --branch "$t"; its release-run read is gh run list --workflow release.yml --limit 1 --json conclusion,headBranch,event,headSha,url
• That row states the match rule explicitly: match the run to $t by headSha against the tag's commit, or ref == refs/tags/$t when present, before declaring a release unhealthy
• That row states that an empty or no-matching-run result is not by itself unhealthy — only a matching run whose conclusion is not success is
• bun run --filter @ultimat3/jobs typecheck is clean (the prior run's gate failed with X_TYPECHECK_FAILED naming packages/jobs/src)
• bun run verify is green end to end

Changes

• .dz/briefs/health.md (edit) — Row the last tag published (line 13) must read a TAG dispatch, not a branch: drop --branch "$t" (release.yml:81-92 refuses any ref that is not refs/tags/v*, so a branch filter can never match) and use gh run list --workflow release.yml --limit 1 --json conclusion,headBranch,event,headSha,url, matched to $t by headSha against the tag's commit (or ref == refs/tags/$t when present), with an empty result explicitly NOT itself unhealthy. The working tree already carries this wording — re-read the line and keep it; edit only if it differs.
• packages/jobs/src/queue-wake.job.test.ts (edit) — Keep the applied SQL_ENQUEUE warmup in the cross-process wake test (the 5-iteration enqueue/delete loop before the SOON_MS measurement). Already in the tree and idiomatic (delete from x_jobs where id = $1::uuid matches jobs-retention.job.test.ts:169). Re-read and leave byte-for-byte unless typecheck fails on it; the prior run's gate failed X_TYPECHECK_FAILED naming packages/jobs/src, so verify this file compiles under the jobs workspace typecheck.

Verification

bun run verify passed (exit 0) in 32m 01s.

Notes

Premises of the request the team disproved:

• Advisory list names packages/jobs/src/warmup.tmp.ts. It does not exist: code_glob packages/jobs/src/*.tmp.ts returned [] and grep for warmup over packages/jobs found nothing. The applied diff is the inline loop at queue-wake.job.test.ts:300-310. Declaring that delete is what blocked the prior run's commit.
• The request says health.md:13 still uses --branch "$t". In the tree it does not: .dz/briefs/health.md:13 already reads --limit 1 --json conclusion,headBranch,event,headSha,url, matches by headSha, and calls an empty list not unhealthy.

---

🤖 Opened by developerz.ai for task tsk_05faa135148a465ac8760fec67efa880 on developerz-ai/ultimate.

Raised automatically by the raw_task_triage lane.

🤖 developerz.ai, an automated agent, not a human — this PR is one slice of a stacked plan.
Dispatched automatically by the raw_task_triage lane
Model MiniMax-M3 on router.ai.developerz.ai via native, funded by account_key
Ran on box hv-dedicated-2
Cost unknown: no list price on record for MiniMax-M3 on router.ai.developerz.ai
Run timeline run_89922c4406d243f3a10af48b9ed1cc03
Part 1/1 of plan rtk_47ad3ca784e44927bf9012e1ae671ffd

[developerz-ai]

Health release-run row: read a tag dispatch, not a branch

• Picked up
• Surveying the repo — 2026-10-20 18:23 UTC
• Working — 2026-10-20 18:32 UTC
• Pull request opened — Health release-run row: read a tag dispatch, not a branch #820 — 2026-10-21 04:42 UTC
• Done

Work landed on #820 — watching it toward merge.

102h13m · 414 tool calls · 270 messages

Sessions: coding run

---

🤖 developerz.ai, an automated agent, not a human — hv-dedicated-2 (developerz-ai[bot]) · model MiniMax-M3 on router.ai.developerz.ai

[developerz-ai]

🟡 Reviewed

2 actionable comment(s) · grounded on your code

_{⏱ 1m 05s wall clock · gpt-6.1-sol via router · 2 model call(s) · 667 output token(s)}

_{🤖 developerz.ai — automated review, running on your model and your box.}

[developerz-ai]

Review summary — 2 file(s), 1 finding(s).

Critical 0 · Major 1 · Minor 0 · Nit 0

Completed review [bot]. The PR updates release-run health guidance and warms the jobs timing test, but the health command requests an unsupported JSON field and still treats a missing matching run as unhealthy.

No findings from: concern-tests.

File	Findings
.dz/briefs/health.md	1 major

1 lower-signal comment(s) suppressed (anti-noise: 1 duplicate).

---

🤖 developerz.ai review — automated, running on your model and your box. What is this?
Reviewed by router/gpt-6.1-sol on box hv-dedicated-2 — 48.9k in / 587 out over 3 calls; cost unknown: no price on record for router/gpt-6.1-sol.

[developerz-ai]

A fix round is resolving this pull request's merge conflict with main. Nothing is needed from you right now.

• Red on this head: no check is failing
• What was done: Fix round tsk_f08aed94491c5bb4f6c884cfa2985afe is running: this pull request conflicts with its base branch main, so GitHub cannot merge it until the conflict is resolved.
• What happens next: It merges the base branch into this branch, resolves the conflicting files and pushes (no force-push); CI and the review then run on the new head. If the conflict cannot be settled within the attempt budget, a person is asked.

_{Status for head f96067545c37. This comment is updated in place.}

🤖 Posted by developerz.ai — an automated agent, not a human.

[developerz-ai]

Review summary — 2 file(s), 2 finding(s).

Critical 0 · Major 1 · Minor 1 · Nit 0

Completed review [bot]: fixes the release query fields and relaxes jobs timing bounds, but the health row still declares nonmatching runs unhealthy, contrary to the requested match-only verdict. Completed: the outbox timing test can pass without waking the worker [bot].

1 non-blocking finding(s) below major (no thread on a re-review — only blocking findings open threads)

• packages/jobs/src/queue-wake.job.test.ts:349 minor · test The worker now polls every 2s while the test waits for the relay to idle, so a remaining poll wait below 1s lets the final assertion pass even if worker wake delivery is broken [bot]. Keep the worker's idle ceiling much longer than the relay setup and measurement, or synchronize its next poll befor…

• CI checks were not shown to the reviewer — GitHub reported no checks on the commit under review, so this review judged the diff without a build result

🤖 developerz.ai review — automated, running on your model and your box. What is this? · Reviewed by router/gpt-6.1-sol on box hv-dedicated-3 — 55k in / 900 out over 3 calls; cost unknown: no price on record for router/gpt-6.1-sol.

[developerz-ai]

Review summary — 1 file(s), 0 finding(s).

Completed [bot]: reviewed the revised queue-wake timing tests; no concrete defects found in the supplied diff. The changes separate the no-wake observation window from loaded-runner wake latency bounds.

🤖 developerz.ai review — automated, running on your model and your box. What is this? · Reviewed by router/gpt-6.1-sol on box hv-dedicated-dev-din — 80.2k in / 711 out over 3 calls; cost unknown: no price on record for router/gpt-6.1-sol.

[developerz-ai]

Review summary — 1 file(s), 0 finding(s).

Completed review [bot]: the health brief now assesses release conclusions only for a matching tag commit and treats absent or nonmatching runs as no verdict. No defects found in this increment.

• CI checks were not shown to the reviewer — GitHub reported no checks on the commit under review, so this review judged the diff without a build result
• test files were not shown to the reviewer — this diff touched no path recognised as a test, so no assertion was available to read the change against

🤖 developerz.ai review — automated, running on your model and your box. What is this? · Reviewed by router/gpt-6.1-sol on box hv-dedicated-1 — 17.5k in / 105 out over 1 call; cost unknown: no price on record for router/gpt-6.1-sol.

[developerz-ai]

Review summary — 2 file(s), 2 finding(s).

Critical 0 · Major 0 · Minor 2 · Nit 0

Completed review [bot]: the release-health row now matches the tag commit and treats absent runs as inconclusive, but still misclassifies an in-progress matching publish as unhealthy. Completed test review [bot]: the outbox latency test can pass without waking the worker.

🤖 developerz.ai review — automated, running on your model and your box. What is this? · Reviewed by router/gpt-6.1-sol on box hv-dedicated-reviewer-1 — 59.1k in / 667 out over 2 calls; cost unknown: no price on record for router/gpt-6.1-sol.