Firebase Tests failed — find root cause and fix #560

Open
opened 2026-06-11 03:04:44 +00:00 by forgejo-actions · 1 comment
forgejo-actions commented 2026-06-11 03:04:44 +00:00 (Migrated from codeberg.org)

Firebase instrumented tests failed in the daily run.

Failed run: https://codeberg.org/guettli/sharedinbox/actions/runs/4657175

Steps to resolve

  1. Find the root cause: Check the test run logs linked above and identify which test(s) failed and why.
  2. Fix if possible: If the failure is caused by a code bug, create a fix. If it is a flaky or infrastructure issue, document the findings.
  3. Close this issue once the root cause is resolved and the tests pass.


Agentloop runs

# Started (UTC) Ended (UTC) Status Tokens in Tokens out
1 2026-06-11 03:04:48 2026-06-11 03:15:01 success 92 18 716
Firebase instrumented tests failed in the daily run. **Failed run:** https://codeberg.org/guettli/sharedinbox/actions/runs/4657175 ## Steps to resolve 1. **Find the root cause**: Check the test run logs linked above and identify which test(s) failed and why. 2. **Fix if possible**: If the failure is caused by a code bug, create a fix. If it is a flaky or infrastructure issue, document the findings. 3. Close this issue once the root cause is resolved and the tests pass. --- --- <!-- agentloop-summary --> **Agentloop runs** | # | Started (UTC) | Ended (UTC) | Status | Tokens in | Tokens out | |---|---------------------|---------------------|---------|-----------|------------| | 1 | 2026-06-11 03:04:48 | 2026-06-11 03:15:01 | success | 92 | 18 716 |
guettlibot commented 2026-06-11 03:15:01 +00:00 (Migrated from codeberg.org)

PR #561 created at https://codeberg.org/guettli/sharedinbox/pulls/561.

Summary

Root cause investigation: Issue #560 links to https://codeberg.org/guettli/sharedinbox/actions/runs/4657175 — a 404. The actual failed Firebase Tests run is #2403 (run_id 4657175). Forgejo's web URL for an Actions run uses run_number, not run_id, but firebase-tests.yml hard-codes ${{ github.run_id }} in the auto-created failure issue. The 3m45s total run-time suggests an early-pipeline failure (Dagger setup, build, or gcloud auth) rather than a Firebase Test Lab device-side failure, but Forgejo renders job logs in JavaScript so WebFetch can't extract them, and I have no other read access to the logs — so I cannot pinpoint the specific failing step inside run #2403 without the working log link.

What I changed: PR #561 replaces ${{ github.run_id }} with ${{ github.run_number }} in .forgejo/workflows/firebase-tests.yml, matching the pattern already used in this workflow's own runner-wait-time API queries (and consistent with deploy.yml, website.yml, ci.yml). After this lands, every future Firebase-tests failure issue ships with a working /actions/runs/<run_number> link, so the next run that fails can actually be diagnosed.

Note for the next agentloop iteration: If today's failure was transient infra (likely, given the duration and the recent fix history around de2b9d2 / #549), the next scheduled Firebase run may pass on its own. If it fails again, the failure issue created against this fix will contain a working link and root cause becomes findable.

agentloop · agent: claude

PR #561 created at https://codeberg.org/guettli/sharedinbox/pulls/561. ## Summary **Root cause investigation:** Issue #560 links to `https://codeberg.org/guettli/sharedinbox/actions/runs/4657175` — a 404. The actual failed Firebase Tests run is `#2403` (run_id `4657175`). Forgejo's web URL for an Actions run uses `run_number`, not `run_id`, but `firebase-tests.yml` hard-codes `${{ github.run_id }}` in the auto-created failure issue. The 3m45s total run-time suggests an early-pipeline failure (Dagger setup, build, or gcloud auth) rather than a Firebase Test Lab device-side failure, but Forgejo renders job logs in JavaScript so WebFetch can't extract them, and I have no other read access to the logs — so I cannot pinpoint the specific failing step inside run #2403 without the working log link. **What I changed:** [PR #561](https://codeberg.org/guettli/sharedinbox/pulls/561) replaces `${{ github.run_id }}` with `${{ github.run_number }}` in `.forgejo/workflows/firebase-tests.yml`, matching the pattern already used in this workflow's own runner-wait-time API queries (and consistent with `deploy.yml`, `website.yml`, `ci.yml`). After this lands, every future Firebase-tests failure issue ships with a working `/actions/runs/<run_number>` link, so the next run that fails can actually be diagnosed. **Note for the next agentloop iteration:** If today's failure was transient infra (likely, given the duration and the recent fix history around `de2b9d2` / `#549`), the next scheduled Firebase run may pass on its own. If it fails again, the failure issue created against this fix will contain a working link and root cause becomes findable. <sub>agentloop · agent: claude</sub>
Sign in to join this conversation.
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: guettli/sharedinbox#560