Repository navigation
fix(deploy): keep skipped runs from cancelling queued deploys - #531
Merged
Merged
Conversation
Every completed Test run starts Deploy to Cloudflare, including pull request runs and cancelled main runs. Their deploy job skips, but they still joined the deploy concurrency group, and a newer pending run replaces the one waiting there. On 2026-10-07 a pull request's run cancelled the queued deploy of #526, so production stayed one commit behind main. Only a successful main push can deploy, so every other run now takes a group of its own. The same day the Artifact signer deploy failed once with Cloudflare API code 10013 and passed on a rerun. Both wrangler deploy paths now retry a transient Cloudflare API fault twice, after 15 and 45 seconds. Any other failure stops the job at once.
harlan-zw
added
the
harlan-agent-review
harlan-github-agent
Bot
added
harlan-agent-review-required
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
🔗 Linked issue
No linked issue.
❓ Type of change
📚 Description
Production is one commit behind main: #526 never deployed. Every completed Test run starts
Deploy to Cloudflare, including pull request runs and cancelled main runs. Those runs skip their deploy job, but they still joined thedeploy-cloudflare-siteconcurrency group. GitHub keeps one pending run per group, and a newer run replaces it. At 02:43:00 UTC on 7 October the deploy for #526 was waiting behind a rerun. At 02:43:20 a pull request's Test run finished, its skipped run took the slot, and the #526 deploy was cancelled. The same thing cancelled a queued deploy at 02:35:56.Only a successful main push can deploy. Every other
workflow_runnow takes a concurrency group of its own, so it can no longer reach the deploy queue. Dispatch runs keep their groups.The rerun was needed because the Artifact signer deploy failed once with Cloudflare API code 10013 ("An unknown error has occurred"). It is the only Cloudflare API deploy failure in the 167 deploys since 14 September, and a rerun passed.
scripts/lib/cloudflare-retry.tsnow retries a transient Cloudflare API fault twice, after 15 and 45 seconds. A transient fault is an API unknown error, a 5xx answer, or a dropped connection. Authentication and configuration errors fail at once. The signer step runs through the newpnpm production:deploy-signer, and the site deploy wraps itswranglercommand with the same retry.📝 Migration
Merging this deploys main, which includes #526.