fix: go to gh runner for broken test

[Flo4604]

What does this PR do?

Switches back to normal GH runner for the test that keeps failing, also only runs docs linting when we change docs....

Type of change

• Bug fix (non-breaking change which fixes an issue)
• Chore (refactoring code, technical debt, workflow improvements)
• Enhancement (small improvements)
• New feature (non-breaking change which adds functionality)
• Breaking change (fix or feature that would cause existing functionality to not work as expected)
• This change requires a documentation update

How should this be tested?

Test go green?

Checklist

Required

• Filled out the "How to test" section in this PR
• Read Contributing Guide
• Self-reviewed my own code
• Commented on my code in hard-to-understand areas
• Ran pnpm build
• Ran pnpm fmt
• Checked for warnings, there are none
• Removed all console.logs
• Merged the latest changes from main onto my branch with git pull origin main
• My changes don't cause any responsiveness issues

Appreciated

• If a UI change was made: Added a screen recording or screenshots to this PR
• Updated the Unkey Docs if changes were necessary

[changeset-bot]

⚠️ No Changeset found

Latest commit: 06a5d94

Merging this PR will not cause a version bump for any packages. If these changes should not result in a new version, you're good to go. If these changes should result in a version bump, you need to add a changeset.

This PR includes no changesets

When changesets are added to this PR, you'll see the packages that this PR includes changesets for and the associated semver types

Click here to learn what changesets are, and how to add one.

Click here if you're a maintainer who wants to add a changeset to this PR

[vercel]

The latest updates on your projects. Learn more about Vercel for GitHub.

Project	Deployment	Preview	Comments	Updated (UTC)
engineering	Ready	Preview	Comment	Oct 21, 2025 5:23pm

1 Skipped Deployment

Project	Deployment	Preview	Comments	Updated (UTC)
dashboard	Ignored	Preview		Oct 21, 2025 5:23pm

[github-actions]

Thank you for following the naming conventions for pull request titles! 🙏

[coderabbitai]

📝 Walkthrough

Walkthrough

Adds a change-detection job to the autofix CI workflow to gate downstream tasks, and standardizes several workflows to use ubuntu-latest runners and the official Docker Buildx action in place of custom Blacksmith runner/actions.

Changes

Cohort / File(s)	Summary
Autofix: change detection gating .github/workflows/autofix.ci.yaml	Added detect_changes job calling a reusable workflow; updated lint_docs to needs: [detect_changes] and added an if condition requiring needs.detect_changes.result == 'success' and docs output
Test workflows: runner and Buildx action updates .github/workflows/job_test_api_local.yaml, .github/workflows/job_test_go_api_local.yaml	Replaced specific Blacksmith runners with ubuntu-latest (original values retained as comments); swapped useblacksmith/setup-docker-builder for docker/setup-buildx-action and commented previous action lines
Release workflow: runner change .github/workflows/release.yaml	Changed goreleaser job runs-on from blacksmith-4vcpu-ubuntu-2404 to ubuntu-latest and added commented original runs-on line

Sequence Diagram(s)

sequenceDiagram
    autonumber
    participant PR as Pull Request
    participant GH as GitHub Actions
    participant Detect as detect_changes (reusable)
    participant Lint as lint_docs job

    PR->>GH: trigger autofix.ci workflow
    GH->>Detect: run detect_changes (reusable workflow)
    alt detect_changes success and docs changed
        Detect-->>GH: result: success, docs_output: true
        GH->>Lint: start lint_docs (needs satisfied + if true)
        Lint-->>GH: lint result
    else detect_changes failure or no docs
        Detect-->>GH: result: failure or docs_output:false
        GH-->>Lint: skip lint_docs (condition not met)
    end

Loading

Estimated code review effort

🎯 3 (Moderate) | ⏱️ ~20 minutes

Possibly related PRs

• fix: Make CI actually CI  #3468 — Also adds/uses a detect_changes step and gates downstream jobs on needs.detect_changes.result == 'success'.
• chore: CI improvements part 1. #3176 — Introduced the reusable job_detect_changes.yaml workflow that this PR invokes.
• ci: use blacksmith to check out their docker stats #3813 — Modified the goreleaser workflow's runs-on setting; directly related to the same release workflow change.

Pre-merge checks and finishing touches

✅ Passed checks (3 passed)

Check name	Status	Explanation
Title Check	✅ Passed	The PR title "fix: go to gh runner for broken test" is directly related to the main changes in the changeset. The pull request primarily involves switching CI runners from custom Blacksmith VMs (blacksmith-8vcpu-ubuntu-2404, blacksmith-16vcpu-ubuntu-2404, blacksmith-4vcpu-ubuntu-2404) to ubuntu-latest across multiple workflow files (.github/workflows/autofix.ci.yaml, job_test_api_local.yaml, job_test_go_api_local.yaml, and release.yaml). The title clearly captures this main infrastructure change. While the PR also includes updates to the Docker Buildx action and a new docs-change detection mechanism, the primary focus on migrating to GitHub's standard runners is well-represented in the title, which is appropriately concise and specific.
Description Check	✅ Passed	The PR description follows the required template structure and includes all major sections: "What does this PR do?" (with a clear summary of the changes), "Type of change" (marked as Bug fix), "How should this be tested?" (though minimal), and the "Checklist" section. The description is mostly complete and directly addresses the key changes—switching to GitHub runners and restricting docs linting to documentation changes. While the testing instructions are brief ("Test go green?") and all checklist items remain unchecked, the description is sufficiently detailed for the nature of the changes (CI/workflow infrastructure updates) and aligns with the template requirements.
Docstring Coverage	✅ Passed	No functions found in the changes. Docstring coverage check skipped.

✨ Finishing touches

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Post copyable unit tests in a comment
• Commit unit tests in branch fix/broken-test

---

📜 Recent review details

Configuration used: CodeRabbit UI

Review profile: CHILL

Plan: Pro

📥 Commits

Reviewing files that changed from the base of the PR and between 632edc7 and 06a5d94.

📒 Files selected for processing (2)

• .github/workflows/job_test_go_api_local.yaml (1 hunks)
• .github/workflows/release.yaml (1 hunks)

🧰 Additional context used

🧠 Learnings (1)

📚 Learning: 2025-08-08T19:52:42.113Z

Learnt from: imeyer
PR: unkeyed/unkey#3765
File: .github/workflows/job_detect_changes.yaml:43-43
Timestamp: 2025-08-08T19:52:42.113Z
Learning: In the unkeyed/unkey repository, the workflows `.github/workflows/job_test_api_local.yaml` and `.github/workflows/job_test_go_api_local.yaml` should keep their Blacksmith self-hosted runners (blacksmith-4vcpu-ubuntu-2404 and blacksmith-8vcpu-ubuntu-2404 respectively) as these are performance-critical test jobs that require dedicated resources.

Applied to files:

• .github/workflows/job_test_go_api_local.yaml

⏰ Context from checks skipped due to timeout of 90000ms. You can increase the timeout in your CodeRabbit configuration to a maximum of 15 minutes (900000ms). (3)

• GitHub Check: Test Go API Local / Test
• GitHub Check: Test API / API Test Local
• GitHub Check: Build / Build

🔇 Additional comments (3)

.github/workflows/job_test_go_api_local.yaml (2)

18-19: Verify Docker action change for compatibility.

The switch from useblacksmith/setup-docker-builder to docker/setup-buildx-action is a significant change. Ensure:

• The official Docker Buildx action provides equivalent functionality for the integration tests (especially Docker Compose setup in lines 14-16).
• No breaking changes in how Docker BuildKit environment variables (lines 28-29) are handled.
• The action version pinning is appropriate (v3.11.1).

Consider running a full test cycle to confirm the new action works correctly with your test environment.

---

9-10: Disregard the review comment; the learning's interpretation of PR #3765 does not match its stated intent.

The retrieved learning claims PR #3765 documented that these workflows "should keep" Blacksmith runners. However, PR #3765's actual commit message states "fix(ci): change less important jobs to gha runners," suggesting the change was intentional and these workflows were classified as "less important."

A more recent commit (06a5d94e, today) deliberately standardizes all workflows to ubuntu-latest with the message "just use normal runners." Both job_test_go_api_local.yaml and job_test_api_local.yaml now consistently use ubuntu-latest, reflecting this architecture decision. The current change aligns with this uniform strategy, not a deviation from documented best practices.

Likely an incorrect or invalid review comment.

.github/workflows/release.yaml (1)

13-14: Verify release workflow performance with ubuntu-latest runner.

The goreleaser job is switching from blacksmith-4vcpu-ubuntu-2404 to ubuntu-latest. This is lower-risk than test workflows (runs less frequently), but binary compilation/release builds can be resource-intensive.

Verify that:

• GoReleaser builds complete successfully on ubuntu-latest without timeout issues.
• Build times are acceptable (note: ubuntu-latest is typically slower than dedicated 4vcpu runners).
• No platform-specific build failures occur.

If build times become problematic, consider reverting to a self-hosted runner for this job specifically.

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands and usage tips.}

[graphite-app]

Graphite Automations

"Notify author when CI fails" took an action on this PR • (10/20/25)

1 teammate was notified to this PR based on Andreas Thomas's automation.

"Post a GIF when PR approved" took an action on this PR • (10/21/25)

1 gif was posted to this PR based on Andreas Thomas's automation.

[coderabbitai]

Actionable comments posted: 1

📜 Review details

Configuration used: CodeRabbit UI

Review profile: CHILL

Plan: Pro

📥 Commits

Reviewing files that changed from the base of the PR and between c764837 and 632edc7.

📒 Files selected for processing (2)

• .github/workflows/autofix.ci.yaml (2 hunks)
• .github/workflows/job_test_api_local.yaml (1 hunks)

🧰 Additional context used

🧠 Learnings (1)

📚 Learning: 2025-08-08T19:52:42.113Z

Learnt from: imeyer
PR: unkeyed/unkey#3765
File: .github/workflows/job_detect_changes.yaml:43-43
Timestamp: 2025-08-08T19:52:42.113Z
Learning: In the unkeyed/unkey repository, the workflows `.github/workflows/job_test_api_local.yaml` and `.github/workflows/job_test_go_api_local.yaml` should keep their Blacksmith self-hosted runners (blacksmith-4vcpu-ubuntu-2404 and blacksmith-8vcpu-ubuntu-2404 respectively) as these are performance-critical test jobs that require dedicated resources.

Applied to files:

• .github/workflows/job_test_api_local.yaml

⏰ Context from checks skipped due to timeout of 90000ms. You can increase the timeout in your CodeRabbit configuration to a maximum of 15 minutes (900000ms). (3)

• GitHub Check: Test API / API Test Local
• GitHub Check: Test Go API Local / Test
• GitHub Check: Build / Build

🔇 Additional comments (3)

.github/workflows/job_test_api_local.yaml (1)

15-16: Docker Buildx action update is reasonable.

The change from useblacksmith/setup-docker-builder to the official docker/setup-buildx-action@v3.11.1 is a sensible modernization that aligns with standard CI practices. The prior action reference is retained as a comment for reference.

.github/workflows/autofix.ci.yaml (2)

42-43: Compound if condition mixing needs and github.event context — verify logic.

The lint_docs if condition combines job outputs from needs.detect_changes with GitHub event context. While this pattern is valid, ensure the logic handles these scenarios correctly:

• If detect_changes is skipped (draft PR or non-pull_request event), the condition should still evaluate correctly
• If detect_changes fails, lint_docs will be skipped (is this intended?)

The current condition will cause lint_docs to be skipped if detect_changes.result is not 'success' (e.g., if it's 'skipped' or 'failure'). Verify this is the intended behavior.

Can you confirm that this gating logic aligns with the expected CI behavior?

---

9-11: All verification points confirmed—the workflow is properly configured.

The job_detect_changes.yaml workflow exists and is correctly implemented:

• ✓ Defines a docs output with proper chaining (jobs.build.outputs.docs ← steps.changes.outputs.docs)
• ✓ Includes path patterns for detecting docs changes (apps/docs/, apps/engineering/, *.md)
• ✓ The lint_docs job gating logic is sound:
  • Checks needs.detect_changes.result == 'success' (handles workflow failure)
  • Checks needs.detect_changes.outputs.docs == 'true' (handles missing output)
  • Conditional logic properly handles draft PRs and non-PR events

No issues found.

[imeyer]

[graphite-app]

[Flo4604]

superseeded by #4138