Skip to content

Fail deploy when /reload-theme can't recover#20

Merged
superterran merged 1 commit into
mainfrom
fix/reload-theme-hard-fail
May 24, 2026
Merged

Fail deploy when /reload-theme can't recover#20
superterran merged 1 commit into
mainfrom
fix/reload-theme-hard-fail

Conversation

@superterran
Copy link
Copy Markdown

Summary

  • Theme reload now retries 3x with exponential backoff (2s, 4s) on 5xx responses and network errors before failing.
  • After retries are exhausted the script returns failure and emits a ::error:: annotation, so the workflow turns red and the failure surfaces in the Actions UI.
  • Previously a 500 from /account/themes/reload printed a warning and the job exited green — so a broken theme ID or upstream outage produced a silent no-op deploy.

Test plan

  • Push a no-op change to a theme path on a deployable branch and confirm a healthy run still exits green.
  • (Optional) Temporarily point MICROBLOG_THEME_ID at an invalid value and confirm the workflow fails with the new ::error:: annotation after 3 attempts.

🤖 Generated with Claude Code

The reload_theme step previously returned success on any non-2xx/302
response, so a 500 from Micro.blog would log a warning and the deploy
job would still exit green — masking real failures (e.g. wrong theme ID
or upstream outage).

Now retries 3x with exponential backoff (2s, 4s) on 5xx and network
errors, then fails hard with a ::error:: annotation so the workflow
turns red and the failure is visible in the Actions UI.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
@superterran superterran merged commit 4e19dac into main May 24, 2026
1 check passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant