CI/CD Guide

Key Takeaways

CI/CD automates code integration, testing, and deployment to reduce manual work and ship faster. The three-tier system—CI (automated build/test), Continuous Delivery (automated to gate), Continuous Deployment (fully automated)—trips up many candidates. Know the deployment strategies (Blue/Green, Canary, Rolling), secrets management, and rollback tactics cold.

What is CI/CD?

CI/CD is the backbone of modern software delivery. It automates the journey from code commit to production, catching bugs early and enabling rapid releases. Teams using CI/CD ship with confidence—broken code never reaches users.

CI vs Continuous Delivery vs Continuous Deployment

Practice	Scope	Manual Steps	When to Use
CI (Continuous Integration)	Merge to mainline frequently; automated build + test on every commit	Deploy step	Bleeding-edge teams, fast iteration cycles
Continuous Delivery	Automated testing and build; manual trigger to production	Deploy approval required	Enterprise, regulated industries (banking, healthcare)
Continuous Deployment	Every passing change ships to prod automatically	None	SaaS, high-confidence teams, rapid feedback needed

Interview Angle

This distinction is a classic gotcha—interviewers love asking “What’s the difference between Continuous Delivery and Continuous Deployment?” The key: Delivery = manual gate, Deployment = fully automatic. Write that down.

The CI/CD Pipeline

A healthy pipeline is structured, measurable, and fast:

1. Source — Code push to mainline branch triggers everything
2. Build — Compile code, resolve dependencies
3. Test — Unit tests (fast, many), integration tests, coverage gates
4. Security Scan — SAST, dependency checks, secrets detection
5. Deploy to Staging — Smoke tests, performance checks in prod-like environment
6. Deploy to Production — Release to users

Each stage is a gate—failure stops the pipeline before bad code ships.

Deployment Strategies

Choosing the right strategy determines how seamlessly your release rolls out and how quickly you can recover from issues.

Strategy	Downtime	Rollback Speed	Best For	Risk Level
Blue/Green	None (instant traffic swap)	Immediate (switch back)	Critical services, zero-downtime requirement	Low
Canary	None (gradual rollout)	Gradual (increase traffic back)	Testing in prod, catching issues early	Medium
Rolling	None (staged updates)	Slow (restart services)	Stateless services, cost-conscious teams	Medium-High

Interview Angle

“When would you use Blue/Green vs Canary?” — Blue/Green for instant rollback needs, Canary for early-warning detection. Blue/Green requires 2x infrastructure; Canary monitors real traffic percentages. Know why you’d pick each.

Popular CI/CD Tools

Tool	Best For	Key Feature
GitHub Actions	GitHub repos, teams already in GitHub ecosystem	Free for public repos, native GitHub integration
GitLab CI/CD	Full DevOps stack, on-premise control	Built-in container registry, GitOps-ready
Jenkins	Enterprise, custom workflows	Open-source, highly extensible via plugins
Azure DevOps	Microsoft shops, enterprise governance	Full pipeline + artifact + repo in one platform

Common Pitfalls & Best Practices

Common Gotchas

Secrets in code — Never hardcode API keys or credentials in repos; use encrypted environment variables or a vault (HashiCorp Vault, AWS Secrets Manager). Weak test pyramid — Too many E2E tests, too few unit tests = slow feedback loop; invert it (many unit, fewer E2E). No rollback plan — You WILL have a bad deploy; know how to roll back faster than you rolled forward.

The Rollback Question

Interview Angle

“Your deploy just crashed production. How do you roll back?” — This is the #1 gotcha in interviews. Have a rehearsed answer: “We use Blue/Green so we instant-swap traffic back to the previous version in seconds. For Canary, we dial traffic back. For Rolling, we restart services with the old image. We always test rollbacks before going live.” Bonus: mention feature flags as a safer escape hatch.

Every team should practice rollbacks in staging regularly. When the pager goes off at 2 AM, you’ll thank past-you for the muscle memory.