CI vs CD vs CD

# ===== Three letters, three commitments =====
#
# CI  Continuous Integration
#     Every commit on every branch is built and tested automatically.
#     If it goes red, fixing it is the team's top priority.
#
# CD  Continuous Delivery
#     Every successful main build produces a deployable artifact.
#     Deploys are one-click (button or chat command), staged via gates.
#
# CD  Continuous Deployment
#     Every successful main build deploys to production automatically.
#     No human gate after green tests; the safety net is automated rollback.

# ===== Pick the right level =====
# Stage 1 — CI only
#   Tests run on every PR + push. Merging needs green tests.
#   Right when: solo or small team, traditional release cycles, manual deploys.

# Stage 2 — Continuous Delivery
#   Main builds an artifact (Docker image, signed bundle).
#   A deploy script promotes artifacts: dev -> staging -> prod.
#   Right when: you ship weekly+, want auditable deploys, have rollback verified.

# Stage 3 — Continuous Deployment
#   Same pipeline, no manual gate.
#   Right when: high test coverage, fast rollback, feature flags for risky changes,
#   on-call coverage during deploy windows.

# ===== What changes in the YAML at each level =====
# CI:
name: ci
on: [push, pull_request]
jobs:
  test:
    runs-on: ubuntu-latest
    steps:
      - uses: actions/checkout@v4
      - uses: actions/setup-node@v4 with: { node-version: 20 }
      - run: npm ci
      - run: npm test

# Continuous Delivery (adds build + push + manual deploy):
  build:
    needs: test
    if: github.ref == 'refs/heads/main'
    runs-on: ubuntu-latest
    steps:
      - uses: actions/checkout@v4
      - run: docker build -t ghcr.io/org/app:${{ github.sha }} .
      - run: docker push  ghcr.io/org/app:${{ github.sha }}
  deploy_staging:
    needs: build
    environment: staging
    runs-on: ubuntu-latest
    steps:
      - run: ./deploy.sh staging ${{ github.sha }}
  deploy_prod:
    needs: deploy_staging
    environment:
      name: production         # requires manual reviewers in repo settings
    runs-on: ubuntu-latest
    steps:
      - run: ./deploy.sh prod ${{ github.sha }}

# Continuous Deployment (removes the manual gate):
  deploy_prod:
    needs: deploy_staging
    runs-on: ubuntu-latest
    steps:
      - run: ./deploy.sh prod ${{ github.sha }}        # auto on green main

# ===== The investments each level needs =====
# CI:    fast tests (<10 min), required status checks on main, branch protection
# CD-1:  build artifacts, environments + secrets, smoke tests post-deploy, rollback runbook
# CD-2:  feature flags, canary / blue-green deploys, automated rollback on SLO breach, on-call

# ===== Patterns to internalise =====
# - You walk the ladder. Skipping straight to Continuous Deployment without flags + rollback ends badly.
# - Promote the SAME artifact through environments; never rebuild for prod.
# - Manual gates exist to absorb risk you have not yet automated away. Replace them with tests over time.
# - The gap between 'green tests' and 'safe to deploy' is closed by feature flags and observability.

# ===== Pitfalls =====
# - Calling deploy-from-CI 'CI/CD' but rebuilding the image at each stage -> non-reproducible
# - 'CD' meaning Continuous Deployment in one room, Continuous Delivery in another -> say which
# - Skipping post-deploy verification because tests are green -> drift between staging and prod surfaces here
# - Pipelines longer than 30 min -> people batch commits to avoid the wait, defeating CI

# CI: merge code often; verify with tests + lints.
# CDelivery: every green build is releasable.
# CDeployment: every green build is auto-released.