ci(cloud-deploy): per-env deploy-job concurrency so prod deploys can't be cancelled mid-flight (#11640)

[NubsCarson]

Closes #11640. [cloud-money] / deploy lane.

Bug (launch-critical)

migrate-db has a job-level cloud-db-migrate-<env> concurrency group (cancel-in-progress:false), but the deploy jobs don't — they inherited only the workflow-level group, which didn't protect the in-flight job. With rapid develop→main promotes, every newer main run preempted the running Deploy API Worker job → no Worker deploy completed since 19:13Z, so prod ran stale code (only Pages/console updated). Evidence on the issue: Deploy API Worker cancelled on 854455d + 1fb1c53×2.

Fix

Per-job, per-env concurrency (job-level groups are repo-wide → serialize the same job across runs, queue never cancel):

• deploy-api → cloud-cf-deploy-api-<production|staging>, cancel-in-progress:false
• deploy-console / deploy-app → …-<pr-N|production|staging>, cancel-in-progress only for PR previews

Distinct group per job so the three deploys still parallelize within a run; PR previews still dedupe.

Verification

Can't run a workflow locally, so: YAML validated (yaml.safe_load parses; all 4 jobs carry concurrency), groups confirmed distinct per job, and the pattern mirrors the already-working migrate-db protection + the workflow-level env resolution. Prove-out is the next promote completing a Worker deploy.

Touches the multi-agent-contested cloud-cf-deploy.yml — flagging rather than self-merging; resolve any conflict to develop's version. cc @standujar (deploy/ops)

[greptile-apps]

Your trial has ended. Reactivate Greptile to resume code reviews.

[coderabbitai]

Important

Review skipped

Auto reviews are disabled on this repository. Please check the settings in the CodeRabbit UI or the .coderabbit.yaml file in this repository. To trigger a single review, invoke the @coderabbitai review command.

⚙️ Run configuration

Configuration used: Organization UI

Review profile: CHILL

Plan: Pro

Run ID: 24dd6878-5970-4381-9228-64e0c5a7219e

You can disable this status message by setting the reviews.review_status to false in the CodeRabbit configuration file.

Use the checkbox below for a quick retry:

• 🔍 Trigger review

✨ Finishing Touches

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Commit unit tests in branch fix/11640-deploy-job-concurrency

---

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.}

[greptile-apps]

Your trial has ended. Reactivate Greptile to resume code reviews.

[lalalune]

Reviewed the workflow-only fix and pushed evidence commit cefa56e9a0.

Validation performed:

python3 YAML parse + job concurrency check
=> migrate-db, deploy-api, deploy-console, deploy-app all have job-level concurrency

migrate-db: cloud-db-migrate-<production|staging>, cancel-in-progress: false
deploy-api: cloud-cf-deploy-api-<production|staging>, cancel-in-progress: false
deploy-console: cloud-cf-deploy-console-<pr-N|production|staging>, PR previews cancel, branch deploys queue
deploy-app: cloud-cf-deploy-app-<pr-N|production|staging>, PR previews cancel, branch deploys queue

git diff --check origin/develop...HEAD
=> clean

I also added .github/issue-evidence/11640-cloud-deploy-job-concurrency.md. Biome ignores that markdown evidence path, so there is no meaningful Biome file check for this PR beyond YAML parse/diff-check.

Manual review: the groups are distinct per deploy job, so API/Console/App can still run in parallel within a single workflow run; job-level groups serialize only the same deploy job across runs. This matches the issue intent and preserves PR preview dedupe for Pages jobs.

[lalalune]

Approved after workflow review and evidence commit cefa56e9a0. The job-level concurrency groups protect production/staging deploy jobs from mid-flight cancellation while preserving PR preview dedupe and intra-run API/Console/App parallelism.

[claude]

Claude encountered an error —— View job

---

I'll analyze this and get back to you.