Why should you care?

Manual releases on Friday cause outages; pipeline enforces same steps every time.

See it live — copy this example

Paste into your MeanVerse project (Angular + Express + MongoDB), then run with ng serve / node / mongosh as noted.

# Pipeline stages (GitHub Actions / Azure DevOps)
# 1 lint -> 2 unit test -> 3 integration test -> 4 docker build
# 5 scan (trivy) -> 6 deploy staging -> 7 smoke test -> 8 deploy prod (approval)
- name: Deploy staging
  if: github.ref == 'refs/heads/main'
  run: kubectl set image deployment/meanverse-api api=ghcr.io/meanverse/api:${{ github.sha }} -n staging
- name: Smoke test
  run: curl -f https://staging-api.meanverse.com/health

What happened?

• Each stage gates next.
• Staging deploy verifies sha image.
• Smoke test catches broken deploy before prod.
• Manual approval for production optional.

Practice next

1. Map branches: PR -> CI, main -> staging, tag -> prod.
2. Automate database migration job in pipeline.
3. Keep rollback: kubectl rollout undo one command.
4. Add blue-green deploy with second deployment.
5. Feature flags decouple deploy from release.

Remember

CI/CD = repeatable path to production. Test and scan before deploy. Staging mirrors prod topology.

Weekly release train

MeanVerse ships 12 PRs merged to main every Tuesday.

Outcome: Pipeline promotes one tested artifact staging -> prod with zero downtime.