Troubleshooting

FixSense Doesn't Comment on My PR

Check installation status

1. Go to your FixSense Dashboard
2. Verify your installation shows as active
3. If not listed:
   • GitHub: Reinstall the GitHub App from Settings → GitHub Apps
   • GitLab / Azure DevOps / Bitbucket: Go to Setup and reconnect your project

Check if the workflow failed

FixSense only analyzes failed CI runs. If your tests passed, no comment will be posted.

Check your monthly limit

If you've reached your monthly analysis quota, FixSense will post a single notification comment instead of individual analyses. Check your dashboard for current usage.

Verify the workflow contains test results

FixSense looks for test framework artifacts (JUnit XML, Playwright JSON, Mochawesome, TRX) in your CI output. Make sure your CI pipeline uploads test result artifacts.

Auto-Fix Not Triggering

Verify the workflow file exists

• GitHub: Ensure .github/workflows/failsense-fix.yml exists in your repository. See Configuration for setup instructions.
• GitLab: Ensure the .gitlab-ci.yml includes a FixSense fix job.
• Azure DevOps: Ensure the pipeline YAML includes the fix step.
• Bitbucket: Ensure bitbucket-pipelines.yml has a fixsense-fix custom pipeline.

Check your AI API key

The ANTHROPIC_API_KEY secret must be set in your CI environment:

• GitHub: Repo → Settings → Secrets and variables → Actions
• GitLab: Settings → CI/CD → Variables
• Azure DevOps: Pipelines → Library → Variable groups
• Bitbucket: Repository settings → Pipelines → Repository variables

Check auto-fix limits

Auto-fixes have separate monthly limits per plan. Check your dashboard to see remaining auto-fix credits.

Analysis Quality Issues

Analysis seems inaccurate

The AI analysis depends on the quality of information available. To improve accuracy:

• Use descriptive test names that explain what the test verifies
• Include meaningful assertion messages in your tests
• Keep tests focused — test one thing per test case
• Enable detailed reporting for maximum log context

Analysis is too generic

If analyses seem generic ("test failed due to an assertion error"), the CI logs may not contain enough context. Try:

1. Add detailed reporters to your test framework config
2. Enable trace recording on failure: use: { trace: 'on-first-retry' }
3. Make sure error messages include expected vs actual values

Dashboard Issues

Dashboard shows no data

• Verify you're signed in with the same account used to connect FixSense
• Check that at least one CI run has failed since installation
• Data may take up to 60 seconds to appear after a CI run

Usage numbers don't match

Analysis counts update in real-time but the dashboard may cache data for a few seconds. Refresh the page to see the latest numbers.

Still Need Help?

If your issue isn't covered here, open an issue on GitHub with your repository name and the specific problem.