[Fleet Monitor] petry-projects/.github-private — .github/workflows/test-dev-lead.yml #1538

Activity

1. don-petry

   added

   health-checkAutomated health check reportAutomated health check report

   dev-leadFor dev-lead agent pickupFor dev-lead agent pickup

   fleet-trackerTracked by the Actions Fleet MonitorTracked by the Actions Fleet Monitor

   on Aug 19, 2026
2. don-petry commented on Aug 19, 2026

   

   don-petry

   on Aug 19, 2026

   CollaboratorAuthor

   More actions

   Dev-Lead Implementation Plan

   Issue: #1538 — [Fleet Monitor] petry-projects/.github-private — .github/workflows/test-dev-lead.yml

   Root cause

   The DEGRADED signal traces to the unit job, where tests/dev-lead/unit/test_advisory_review_gate.bats
   fails intermittently. In every failing run the same cluster fails (tests 31/33/35/38/39/40/45) while
   every "expect status 1" gate test passes. That fingerprint — including test 45 (which only asserts the
   gate's output contains RATE_LIMITED, regardless of exit code) producing no output — means the gate
   script is exit 1-ing at source time.

   scripts/lib/advisory-review-gate.sh sources the reviewer-source registry and, if
   reviewer_sources_advisory_gate_logins (or reviewer_sources_logins) fails, runs exit 1 (lines 42 and
   212). Because the gate is sourced, that exit kills the caller — the whole bats invocation for that
   test — with no output and status 1. A sourced library must never exit its caller on a recoverable read
   error. The gate already carries a built-in fallback bot list for the "registry file absent" case, but
   hard-exits instead of using it when the read fails.

   Confirmed deterministically: sourcing the gate with REVIEWER_SOURCES_MANIFEST=/nonexistent reproduces the
   exact CI failure signature (no output, status 1).

   Scope

   Make the advisory gate resilient: on a registry-read failure, fall back to the existing built-in bot lists
   (with a warning) instead of exit 1, so a transient/edge read failure can neither strand PR reviews in
   production nor fail the unit suite in CI.

   Implementation Checklist

   • scripts/lib/advisory-review-gate.sh: replace the exit 1 on reviewer_sources_advisory_gate_logins
     failure with a graceful fallback to the built-in ADVISORY_BOTS set
   • scripts/lib/advisory-review-gate.sh: replace the exit 1 on reviewer_sources_logins failure with a
     graceful fallback to the built-in RATE_LIMIT_NOTICE_BOTS set

   Tests to Write

   • tests/dev-lead/unit/test_advisory_review_gate.bats: sourcing survives an unreadable registry (no
     exit, ADVISORY_BOTS populated from fallback)
   • tests/dev-lead/unit/test_advisory_review_gate.bats: check_advisory_reviews still runs (applies the
     head-age timeout) when the registry is unreadable

   Test Command

   bats tests/dev-lead/unit/ (full unit job), plus bash .dev-lead/scripts/dev-lead-lint.sh
3. don-petry commented on Aug 19, 2026

   

   don-petry

   on Aug 19, 2026

   CollaboratorAuthor

   More actions

   Dev-Lead: Implementation Complete — PR #1540

   The implementation for issue #1538 is durable: commits are pushed and a pull request is open. This record is backed by verifiable artifacts you can confirm in one click.

   • Pull request: feat: implement issue #1538 — [Fleet Monitor] petry-projects/.github-private — .github/workflows/test-dev-lead.yml #1540
   • Head commit: 5695ec8d390e5b3d7a9bfdebcd0f95e1d410dc57
   • Run: https://github.com/petry-projects/.github-private/actions/runs/32227461947

   Review happens on the PR. Acceptance criteria and test results are evidenced by the PR's diff and CI checks — not asserted here. If a later stage fails, this claim is retracted in place.
4. don-petry

   closed this as completedin #1540on Aug 19, 2026
5. don-petry

   added a commit that references this issue on Aug 19, 2026

   feat: implement issue #1538 — [Fleet Monitor] petry-projects/.github-…

   ...

   99d1875