Add manual workflow to reset failed builds at max retries#279
Add manual workflow to reset failed builds at max retries#279frostebite wants to merge 1 commit intomainfrom
Conversation
Manually-triggered workflow that calls the resetFailedBuilds backend endpoint. Can reset a single build by ID or bulk-reset all builds stuck at max failure count. Follows the same pattern as clean-up-stuck-builds.yml. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
|
|
📝 WalkthroughWalkthroughA new GitHub Actions workflow is added for resetting failed builds via manual dispatch. The workflow requires a confirmation input and optionally accepts a build ID parameter to reset either a specific build or all maxed-out builds via an authenticated API call. Changes
Sequence DiagramsequenceDiagram
actor User
participant GitHub Actions
participant Cloud Function
User->>GitHub Actions: Trigger workflow_dispatch
GitHub Actions->>GitHub Actions: Validate confirm='yes'
alt confirm != 'yes'
GitHub Actions->>User: Abort workflow
else confirm == 'yes'
alt buildId provided
GitHub Actions->>GitHub Actions: Build payload {buildId}
else buildId empty
GitHub Actions->>GitHub Actions: Build payload {}
end
GitHub Actions->>Cloud Function: POST with VERSIONING_TOKEN
Cloud Function->>Cloud Function: Process reset request
Cloud Function-->>GitHub Actions: HTTP Response
alt Status == 200
GitHub Actions->>GitHub Actions: Parse response JSON
GitHub Actions->>User: Emit success notice
else Status != 200
GitHub Actions->>User: Exit non-zero / Fail job
end
end
Estimated code review effort🎯 2 (Simple) | ⏱️ ~10 minutes Poem
🚥 Pre-merge checks | ✅ 3✅ Passed checks (3 passed)
✏️ Tip: You can configure your own custom pre-merge checks in the settings. ✨ Finishing Touches🧪 Generate unit tests (beta)
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. Comment |
![coderabbitai[bot]](https://avatars.githubusercontent.com/in/347564?s=60&v=4){kind=small}
There was a problem hiding this comment.
Actionable comments posted: 1
🧹 Nitpick comments (1)
.github/workflows/reset-failed-builds.yml (1)
33-37: Harden the backend call with timeout and transport-error handling.Recommend adding
--show-error,--connect-timeout, and--max-time(and handling curl non-zero exit) to avoid hanging/manual triage on transient network failures.Proposed resiliency update
- RESPONSE=$(curl -s -w "\n%{http_code}" -X POST \ + RESPONSE=$(curl -sS --connect-timeout 10 --max-time 60 -w "\n%{http_code}" -X POST \ -H "Authorization: Bearer ${{ secrets.VERSIONING_TOKEN }}" \ -H "Content-Type: application/json" \ -d "$PAYLOAD" \ - "https://europe-west3-unity-ci-versions.cloudfunctions.net/resetFailedBuilds") + "https://europe-west3-unity-ci-versions.cloudfunctions.net/resetFailedBuilds") || { + echo "::error::Request failed before receiving an HTTP response" + exit 1 + }🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In @.github/workflows/reset-failed-builds.yml around lines 33 - 37, The curl call that assigns RESPONSE should be hardened: add --show-error, --connect-timeout (e.g. 10) and --max-time (e.g. 30) flags to the curl invocation that posts to "https://.../resetFailedBuilds" and capture curl's exit status; after running curl (the RESPONSE assignment for the POST to resetFailedBuilds) check the exit code ($?) and if non-zero log an error and fail the job (or retry) instead of proceeding with a possibly empty/hanging RESPONSE so transport errors/timeouts are surfaced and handled.
🤖 Prompt for all review comments with AI agents
Verify each finding against the current code and only fix it if needed.
Inline comments:
In @.github/workflows/reset-failed-builds.yml:
- Line 27: The PAYLOAD construction currently interpolates BUILD_ID directly
into a JSON string which will break for quotes/backslashes; change the logic
that sets PAYLOAD (the PAYLOAD variable that uses BUILD_ID) to build the JSON
safely using jq --arg (e.g., echo '{}' | jq --arg buildId "$BUILD_ID" '.buildId
= $buildId') or an equivalent jq command so BUILD_ID is properly escaped and
valid JSON is produced.
---
Nitpick comments:
In @.github/workflows/reset-failed-builds.yml:
- Around line 33-37: The curl call that assigns RESPONSE should be hardened: add
--show-error, --connect-timeout (e.g. 10) and --max-time (e.g. 30) flags to the
curl invocation that posts to "https://.../resetFailedBuilds" and capture curl's
exit status; after running curl (the RESPONSE assignment for the POST to
resetFailedBuilds) check the exit code ($?) and if non-zero log an error and
fail the job (or retry) instead of proceeding with a possibly empty/hanging
RESPONSE so transport errors/timeouts are surfaced and handled.
🪄 Autofix (Beta)
Fix all unresolved CodeRabbit comments on this PR:
- Push a commit to this branch (recommended)
- Create a new PR with the fixes
ℹ️ Review info
⚙️ Run configuration
Configuration used: defaults
Review profile: CHILL
Plan: Pro
Run ID: 1a5a4522-a9f8-43cd-b917-378b74495bab
📒 Files selected for processing (1)
.github/workflows/reset-failed-builds.yml
|
|
||
| if [ -n "$BUILD_ID" ]; then | ||
| echo "Resetting failure count for build: $BUILD_ID" | ||
| PAYLOAD="{\"buildId\": \"$BUILD_ID\"}" |
There was a problem hiding this comment.
🧩 Analysis chain
🏁 Script executed:
#!/bin/bash
# Verifies current interpolation can produce invalid JSON.
BUILD_ID='editor-linux-"bad"-id'
PAYLOAD="{\"buildId\": \"$BUILD_ID\"}"
echo "$PAYLOAD" | jq . >/dev/nullRepository: game-ci/docker
Length of output: 115
🏁 Script executed:
cat .github/workflows/reset-failed-builds.ymlRepository: game-ci/docker
Length of output: 1638
🏁 Script executed:
#!/bin/bash
# Verify the proposed jq fix works with problematic BUILD_ID
BUILD_ID='editor-linux-"bad"-id'
PAYLOAD=$(jq -n --arg buildId "$BUILD_ID" '{buildId: $buildId}')
echo "$PAYLOAD" | jq . >/dev/null && echo "Valid JSON: $PAYLOAD"Repository: game-ci/docker
Length of output: 112
Safely escape buildId when constructing JSON payload.
The code directly interpolates user input (github.event.inputs.buildId) into the JSON string, which will produce invalid JSON if the input contains quotes or backslashes. Use jq --arg to safely construct the payload:
Proposed fix
if [ -n "$BUILD_ID" ]; then
echo "Resetting failure count for build: $BUILD_ID"
- PAYLOAD="{\"buildId\": \"$BUILD_ID\"}"
+ PAYLOAD=$(jq -n --arg buildId "$BUILD_ID" '{buildId: $buildId}')
else
echo "Resetting ALL failed builds at max retries..."
PAYLOAD="{}"
fi📝 Committable suggestion
‼️ IMPORTANT
Carefully review the code before committing. Ensure that it accurately replaces the highlighted code, contains no missing lines, and has no issues with indentation. Thoroughly test & benchmark the code to ensure it meets the requirements.
| PAYLOAD="{\"buildId\": \"$BUILD_ID\"}" | |
| if [ -n "$BUILD_ID" ]; then | |
| echo "Resetting failure count for build: $BUILD_ID" | |
| PAYLOAD=$(jq -n --arg buildId "$BUILD_ID" '{buildId: $buildId}') | |
| else | |
| echo "Resetting ALL failed builds at max retries..." | |
| PAYLOAD="{}" | |
| fi |
🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed.
In @.github/workflows/reset-failed-builds.yml at line 27, The PAYLOAD
construction currently interpolates BUILD_ID directly into a JSON string which
will break for quotes/backslashes; change the logic that sets PAYLOAD (the
PAYLOAD variable that uses BUILD_ID) to build the JSON safely using jq --arg
(e.g., echo '{}' | jq --arg buildId "$BUILD_ID" '.buildId = $buildId') or an
equivalent jq command so BUILD_ID is properly escaped and valid JSON is
produced.
webbertakken
left a comment
webbertakken
left a comment
There was a problem hiding this comment.
Personally I don't think we need this if we're going to make the UI more user friendly. The GitHub UI requires you to know the exact parameters etc, whereas in our UI we have all the parameters ready. So in my opinion upgrading https://game.ci/docs/docker/versions/ to become really great would be preferable.
we do have this, I usually go for a manual workflow too, but I am not opposed to desisting with it |
|
Closing per preference: all build management tools via documentation admin UI, not workflow-based interventions. |
|
|
2 participants
Summary
Add a
workflow_dispatchworkflow that calls theresetFailedBuildsbackend endpoint (game-ci/versioning-backend#85). Maintainers can trigger it from the Actions tab to reset builds stuck at max failure count.buildIdempty to reset all maxed-out buildsclean-up-stuck-builds.yml)VERSIONING_TOKENsecretDepends on
resetFailedBuildsendpointTest plan
failureCount >= 15🤖 Generated with Claude Code
Summary by CodeRabbit