簡介
CI/CD Pipeline 擁有 10 多個微服務和 5 個以上微前端，是獨立部署的骨幹。本文設計了一個端到端的管道：從提交到生產。
1. CI/CD 架構
┌─────────────────────────────────────────────────┐
│ Developer pushes code │
└──────────────────────┬──────────────────────────┘
│
┌──────────────────────┴──────────────────────────┐
│ CI Pipeline (GitHub Actions) │
│ ┌─────────┐ ┌──────────┐ ┌───────────────────┐ │
│ │ Lint │ │ Build │ │ Test │ │
│ │ & Type │→│ & Compile│→│ Unit + Integration│ │
│ │ Check │ │ │ │ + Contract │ │
│ └─────────┘ └──────────┘ └───────────────────┘ │
│ │ │
│ ┌──────────────┴──────────┐ │
│ │ Docker Build & Push │ │
│ │ → Container Registry │ │
│ └──────────────┬──────────┘ │
└────────────────────────────────────┼────────────┘
│
┌────────────────────────────────────┴────────────┐
│ CD Pipeline (ArgoCD - GitOps) │
│ │
│ ┌──────────┐ ┌──────────┐ ┌────────────────┐ │
│ │ Staging │→ │ Smoke │→ │ Production │ │
│ │ Deploy │ │ Tests │ │ Deploy │ │
│ │ │ │ │ │ (Progressive) │ │
│ └──────────┘ └──────────┘ └────────────────┘ │
└─────────────────────────────────────────────────┘
2. CI 管道（GitHub Actions）
2.1 微服務管道
# .github/workflows/service-ci.yml
name: Service CI
on:
push:
paths:
- 'services/product-service/**'
- 'packages/shared-types/**'
jobs:
ci:
runs-on: ubuntu-latest
steps:
- uses: actions/checkout@v4
- name: Lint & Type Check
run: turbo lint typecheck --filter=product-service
- name: Unit Tests
run: turbo test --filter=product-service
- name: Integration Tests
run: turbo test:integration --filter=product-service
services:
postgres:
image: postgres:16
env:
POSTGRES_DB: test
- name: Contract Verification
run: turbo test:pact --filter=product-service
- name: Docker Build & Push
run: |
docker build -t ghcr.io/company/product-service:${{ github.sha }} \
-f services/product-service/Dockerfile .
docker push ghcr.io/company/product-service:${{ github.sha }}
2.2 微前端管道
# .github/workflows/mfe-ci.yml
name: MFE CI
on:
push:
paths:
- 'apps/product-mfe/**'
- 'packages/shared-ui/**'
jobs:
ci:
runs-on: ubuntu-latest
steps:
- uses: actions/checkout@v4
- name: Install & Build
run: turbo build --filter=product-mfe...
- name: Test
run: turbo test --filter=product-mfe
- name: Visual Regression
uses: chromatic/action@v1
with:
projectToken: ${{ secrets.CHROMATIC_TOKEN }}
- name: Deploy to CDN
run: |
aws s3 sync apps/product-mfe/dist/ \
s3://mfe-cdn/product-mfe/${{ github.sha }}/
aws cloudfront create-invalidation \
--distribution-id $CF_ID \
--paths "/product-mfe/*"
3. Docker 多階段構建
# services/product-service/Dockerfile
# Stage 1: Build
FROM node:20-alpine AS builder
WORKDIR /app
COPY package*.json ./
RUN npm ci --production=false
COPY . .
RUN npm run build
# Stage 2: Production
FROM node:20-alpine AS runner
WORKDIR /app
RUN addgroup -S appgroup && adduser -S appuser -G appgroup
COPY --from=builder /app/dist ./dist
COPY --from=builder /app/node_modules ./node_modules
COPY --from=builder /app/package.json ./
USER appuser
EXPOSE 8080
CMD ["node", "dist/main.js"]
4.GitOps 與 ArgoCD
4.1 GitOps 工作流程
Developer pushes code → CI builds image → CI updates manifests repo
│
ArgoCD watches manifests repo ──────────────────────┘
ArgoCD syncs K8s cluster to match manifests
4.2 自訂疊加
# k8s/overlays/production/product-service.yaml
apiVersion: apps/v1
kind: Deployment
metadata:
name: product-service
spec:
replicas: 3
template:
spec:
containers:
- name: product-service
image: ghcr.io/company/product-service:abc123
resources:
requests:
cpu: 100m
memory: 256Mi
limits:
cpu: 500m
memory: 512Mi
5. 平行管道
Commit changes to:
├── services/product-service/ → Product Service Pipeline
├── apps/product-mfe/ → Product MFE Pipeline
└── packages/shared-ui/ → Shared UI Pipeline
↓
Triggers dependent:
→ Product MFE Pipeline
→ Cart MFE Pipeline
→ All MFE consumers
渦輪雷波 --filter=...[origin/main] 自動偵測受影響的包。
6. 管道最佳實踐
| 實踐 | 描述 |
|---|---|
| 快速失敗 | Lint → 單元 → 整合（首先需要一點時間） |
| 並行工作 | Lint + 並行測試運行 |
| 快取 | npm 快取、Docker 層快取、Turborepo 遠端快取 |
| 僅受影響 | 僅建置/測試包變更 |
| 不可變的工件 | 使用 git SHA 標記影像 |
| GitOps | 單獨的設定儲存庫，ArgoCD 同步 |
總結
- CI：Lint → 建置 → 測試 → Docker → 推送（每個服務/MFE）
- CD：帶有 ArgoCD 的 GitOps，聲明性 K8s 清單
- Turborepo：影響檢測，僅更改建置包
- Docker 多階段：小型、安全的生產映像
- 並行管道：每個服務/MFE都有自己的管道
下一篇文章： 第 25 課：部署策略 — 藍/綠、金絲雀與滾動