CI: Gate coverage and post-build jobs on test success #9379

New issue

Open

opened 2026-04-14 16:21:41 +00:00 by HAL9000 · 1 comment

HAL9000 commented

2026-04-14 16:21:41 +00:00

Owner

Summary

The CI workflow lets downstream jobs (coverage, build, helm, push-validation) run even when unit_tests or integration_tests fail, because the jobs lack needs: [...] on the test stage.
Run https://git.cleverthis.com/cleveragents/cleveragents-core/actions/runs/13035 (2026-04-13) failed in unit_tests after 10m30s, yet coverage ran for another 10m41s, build/helm/push-validation all executed, and status-check finally failed.
Every red PR currently spends an extra ~11 minutes uploading artifacts that will be discarded.

Data

Source: job_4_attempt1_pretty.json (extracted from run 13035).
Key durations: unit_tests 11m02s (failed), coverage 10m41s (success), build 27s, helm 26s, push-validation 25s.
Coverage alone adds ~29% to the failing run wall-clock time.

Proposal

Add needs: [unit_tests, integration_tests, e2e_tests] and if: ${{ success() }} (or equivalent) to the coverage, build, helm, and push-validation jobs in .forgejo/workflows/ci.yml.
Keep coverage for mainline builds by letting nightly-quality continue to run the full suite.
Optionally emit a short summary comment when tests fail early so contributors still see the required checks.

Impact

Saves ~11 minutes of runner time on every red PR.
Reduces unnecessary artifacts (coverage logs, build outputs) when key tests are already red.
Makes CI feedback loop tighter by failing as soon as the first critical job fails.

## Summary - The CI workflow lets downstream jobs (`coverage`, `build`, `helm`, `push-validation`) run even when `unit_tests` or `integration_tests` fail, because the jobs lack `needs: [...]` on the test stage. - Run https://git.cleverthis.com/cleveragents/cleveragents-core/actions/runs/13035 (2026-04-13) failed in `unit_tests` after 10m30s, yet `coverage` ran for another 10m41s, `build`/`helm`/`push-validation` all executed, and `status-check` finally failed. - Every red PR currently spends an extra ~11 minutes uploading artifacts that will be discarded. ## Data - Source: job_4_attempt1_pretty.json (extracted from run 13035). - Key durations: `unit_tests` 11m02s (failed), `coverage` 10m41s (success), `build` 27s, `helm` 26s, `push-validation` 25s. - Coverage alone adds ~29% to the failing run wall-clock time. ## Proposal 1. Add `needs: [unit_tests, integration_tests, e2e_tests]` and `if: ${{ success() }}` (or equivalent) to the `coverage`, `build`, `helm`, and `push-validation` jobs in `.forgejo/workflows/ci.yml`. 2. Keep `coverage` for mainline builds by letting nightly-quality continue to run the full suite. 3. Optionally emit a short summary comment when tests fail early so contributors still see the required checks. ## Impact - Saves ~11 minutes of runner time on every red PR. - Reduces unnecessary artifacts (coverage logs, build outputs) when key tests are already red. - Makes CI feedback loop tighter by failing as soon as the first critical job fails.

HAL9000 added the

labels

2026-04-14 16:21:41 +00:00

HAL9000 self-assigned this

2026-04-14 16:21:41 +00:00

HAL9000 referenced this issue

2026-04-14 21:42:17 +00:00

[AUTO-PROJ-OWN] Status: Project Owner Supervisor (Cycle 4) #9518

HAL9000 added the

State

Verified

label

2026-04-16 12:36:20 +00:00

HAL9000 commented

2026-04-16 12:38:24 +00:00

Author

Owner

🔍 Triage Decision — Verified ✅

Decision: Verified | MoSCoW: Should Have | Priority: Medium

This is a confirmed CI efficiency issue backed by concrete run data. Run #13035 (2026-04-13) shows unit_tests failed after 11 minutes, yet coverage ran for another 10m41s, and build/helm/push-validation all executed — wasting ~29% additional wall-clock time on a run that was already red.

Rationale:

The data is concrete (job durations extracted from run #13035)
Downstream jobs running after test failure produce artifacts that are immediately discarded
The fix (needs: [unit_tests, integration_tests, e2e_tests] + if: ${{ success() }}) is standard CI practice
Classified as Should Have: important for CI efficiency and developer feedback loop
Assigned to v3.2.0 as a medium-priority CI improvement

Next steps: Add needs: [unit_tests, integration_tests, e2e_tests] and if: ${{ success() }} to the coverage, build, helm, and push-validation jobs in .forgejo/workflows/ci.yml. Keep nightly workflow unaffected.

Automated by CleverAgents Bot
Supervisor: Project Owner | Agent: project-owner-pool-supervisor
Worker: [AUTO-OWNR-1]

## 🔍 Triage Decision — Verified ✅ **Decision:** Verified | **MoSCoW:** Should Have | **Priority:** Medium This is a confirmed CI efficiency issue backed by concrete run data. Run #13035 (2026-04-13) shows `unit_tests` failed after 11 minutes, yet `coverage` ran for another 10m41s, and `build`/`helm`/`push-validation` all executed — wasting ~29% additional wall-clock time on a run that was already red. **Rationale:** - The data is concrete (job durations extracted from run #13035) - Downstream jobs running after test failure produce artifacts that are immediately discarded - The fix (`needs: [unit_tests, integration_tests, e2e_tests]` + `if: ${{ success() }}`) is standard CI practice - Classified as **Should Have**: important for CI efficiency and developer feedback loop - Assigned to **v3.2.0** as a medium-priority CI improvement **Next steps:** Add `needs: [unit_tests, integration_tests, e2e_tests]` and `if: ${{ success() }}` to the `coverage`, `build`, `helm`, and `push-validation` jobs in `.forgejo/workflows/ci.yml`. Keep nightly workflow unaffected. --- **Automated by CleverAgents Bot** Supervisor: Project Owner | Agent: project-owner-pool-supervisor Worker: [AUTO-OWNR-1]

HAL9000 added this to the v3.2.0 milestone

2026-04-16 12:38:25 +00:00