CI/CD RELEASE CHECKLIST Anrotex | https://www.anrotex.com/blog/cicd-best-practices Save one copy per release. Fill in the details below, link to evidence, and mark each check when verified. For a check that does not apply, record why. If a required check fails, hold the release and agree the next action. RELEASE DETAILS Application / service: Repository: Target environment: Commit SHA: Artifact version / digest: Release owner: Reviewer: Planned deployment time and timezone: Change summary: Pipeline run / release record: 1. VERIFY THE CHANGE [ ] Required checks pass for the exact commit and artifact being released. [ ] Critical user journeys pass in the agreed test environment. [ ] Runtime settings, dependencies and destination match this release. [ ] The artifact contains the expected files and no credentials. Evidence / unresolved issues: 2. CHECK DEPLOYMENT ACCESS AND ORDER [ ] The intended branch, checks and approvals control production deployment. [ ] Jobs have only the permissions and credentials they need. [ ] Untrusted pull requests cannot access production secrets or runners. [ ] Overlapping runs cannot interrupt this deployment or replace it with an older release. Database changes have an agreed execution order. Evidence / unresolved issues: 3. PREPARE RECOVERY [ ] A known-good release is available for the agreed recovery period. [ ] The recovery version works with the current schema and configuration. [ ] Recovery steps have been rehearsed in the agreed test environment. [ ] Data recovery is covered separately where needed; reverting application code will not undo writes or restore deleted data. Recovery procedure / runbook: Known-good artifact and configuration: Schema compatibility / migration plan: Backup and data recovery plan, where required: Recovery owner and contact route: 4. AGREE THE HEALTH CHECKS BEFORE DEPLOYING [ ] A named person can stop the rollout and carry out the recovery plan. [ ] Baselines, limits and an observation window are agreed for this workload. [ ] Checks include a critical user journey and relevant background work. Critical journey and expected result: Error-rate baseline / stop threshold: Latency baseline / stop threshold: Queue, saturation or other workload signal / stop threshold: Dashboard / alert links: Observation duration and reason: Decision if a limit is exceeded: Release decision (proceed / hold), reviewer and time: 5. VERIFY AFTER DEPLOYMENT [ ] Confirm the running commit, artifact, environment and configuration. [ ] Complete the smoke checks and watch health for the agreed window. [ ] Check migrations, background jobs and downstream dependencies. [ ] Record the outcome and hand over any remaining work to a named owner. Deployment started / finished: Observed health and smoke-check evidence: Outcome (completed / stopped / recovered / further work needed): Recovery action and time, if used: Follow-up action / owner / due date: Final reviewer and time: USING THE NODE.JS WORKFLOW Download: https://www.anrotex.com/downloads/node-ci.yml Setup: https://www.anrotex.com/blog/cicd-best-practices#github-actions The workflow builds and tests a Node.js 24 project. It expects a committed package-lock.json, npm scripts named lint, test:ci and build, and output in dist/. Review these settings for your application before saving it as .github/workflows/node-ci.yml. It does not deploy to production. Configure deployment checks and access separately, then verify a failing test blocks release before relying on the workflow. DISCUSS YOUR PIPELINE https://www.anrotex.com/contact?focus=cicd sales@anrotex.com