sethdford CI/CD Pipeline Design
Automating build, test, and deployment through continuous integration and deployment.
Context
You are designing a CI/CD pipeline. Define stages, dependencies, and success criteria.
Domain Context
- Build: Compile, package; generate artifacts
- Test: Run unit, integration, E2E tests; quality gates
- Deploy: Canary, blue-green, rolling; minimize downtime
- Artifacts: Build outputs; cache for speed
- Secrets: Never in code; use secrets manager
Instructions
- Identify Stages: Build, test, deploy, monitor
- Define Triggers: On commit? Pull request? Manual?
- Parallelize: Run independent jobs in parallel
- Cache: Cache dependencies, build outputs for speed
- Gate: Quality gates; don't deploy if tests fail
- Secrets: Use secrets manager; never in logs
- Artifacts: Store and reuse build artifacts
Anti-Patterns
- Slow pipelines (>15 min); too long, feedback is slow
- Secrets in code or logs; instant compromise
- No quality gates; broken code gets deployed
- Parallelization ignored; sequential is slow
- No caching; rebuild everything every time
Further Reading
- Google Cloud CI/CD Best Practices
- GitHub Actions documentation
- GitLab CI/CD documentation