Argo Rollouts #

kubectl create namespace argo-rollouts
kubectl apply -n argo-rollouts \
  -f https://github.com/argoproj/argo-rollouts/releases/latest/download/install.yaml

# macOS
brew install argo-rollouts

# 或从 GitHub Releases 下载
# https://github.com/argoproj/argo-rollouts/releases/latest

# 验证安装
kubectl argo rollouts version

# 启动 Dashboard
kubectl argo rollouts dashboard

# 默认监听 http://localhost:3100

apiVersion: argoproj.io/v1alpha1
kind: Rollout
metadata:
  name: rollouts-demo
spec:
  replicas: 5
  strategy:
    canary:
      steps:
        - setWeight: 20
        - pause: {}
        - setWeight: 40
        - pause: { duration: 10s }
        - setWeight: 60
        - pause: { duration: 10s }
        - setWeight: 80
        - pause: { duration: 10s }
  selector:
    matchLabels:
      app: rollouts-demo
  template:
    metadata:
      labels:
        app: rollouts-demo
    spec:
      containers:
        - name: rollouts-demo
          image: argoproj/rollouts-demo:blue
          ports:
            - containerPort: 8080

# 查看 Rollout 状态
kubectl argo rollouts get rollout rollouts-demo --watch

# 更新镜像触发滚动
kubectl argo rollouts set image rollouts-demo \
  rollouts-demo=argoproj/rollouts-demo:yellow

# 手动推广（解除 pause）
kubectl argo rollouts promote rollouts-demo

# 中止并回滚
kubectl argo rollouts abort rollouts-demo

# 重试已中止的 Rollout
kubectl argo rollouts retry rollouts-demo

# 重启所有 Pod
kubectl argo rollouts restart rollouts-demo

apiVersion: argoproj.io/v1alpha1
kind: Rollout
metadata:
  name: bluegreen-demo
spec:
  replicas: 3
  strategy:
    blueGreen:
      activeService: active-service        # 生产流量 Service
      previewService: preview-service      # 预览 Service（仅测试用）
      autoPromotionEnabled: false          # 禁用自动推广（需手动审批）
      autoPromotionSeconds: 30             # 或：就绪后 30s 自动推广
      previewReplicaCount: 1               # 预览阶段只运行 1 个副本
      scaleDownDelaySeconds: 30            # 切换后延迟 30s 缩容旧版本
      prePromotionAnalysis:                # 切换前分析
        templates:
          - templateName: smoke-tests
      postPromotionAnalysis:               # 切换后分析
        templates:
          - templateName: success-rate
  selector:
    matchLabels:
      app: bluegreen-demo
  template:
    metadata:
      labels:
        app: bluegreen-demo
    spec:
      containers:
        - name: app
          image: myapp:v1

apiVersion: argoproj.io/v1alpha1
kind: Rollout
metadata:
  name: canary-demo
spec:
  replicas: 10
  strategy:
    canary:
      steps:
        - setWeight: 10        # 10% 流量 → canary (1/10 pods)
        - pause: {}             # 等待人工审批
        - setWeight: 30        # 30% 流量 → canary (3/10 pods)
        - pause: { duration: 5m }  # 等待 5 分钟
        - setWeight: 50        # 50%
        - pause: { duration: 5m }
        - setWeight: 100       # 全量
  selector:
    matchLabels:
      app: canary-demo
  template:
    metadata:
      labels:
        app: canary-demo
    spec:
      containers:
        - name: app
          image: myapp:v2

apiVersion: argoproj.io/v1alpha1
kind: Rollout
metadata:
  name: canary-traffic-demo
spec:
  replicas: 5
  strategy:
    canary:
      canaryService: canary-service     # Canary Service
      stableService: stable-service     # Stable Service
      trafficRouting:
        istio:                          # 使用 Istio 流量管理
          virtualService:
            name: myapp-vsvc
            routes:
              - primary
      steps:
        - setWeight: 5
        - pause: {}
        - setWeight: 20
        - analysis:                     # 步骤间插入分析
            templates:
              - templateName: success-rate
        - setWeight: 50
        - pause: { duration: 10m }
        - setWeight: 100
  selector:
    matchLabels:
      app: canary-traffic-demo
  template:
    metadata:
      labels:
        app: canary-traffic-demo
    spec:
      containers:
        - name: app
          image: myapp:v2

strategy:
  canary:
    canaryService: canary-service
    stableService: stable-service
    trafficRouting:
      managedRoutes:              # 管理路由的优先级
        - name: set-header-1
        - name: mirror-route
      istio:
        virtualService:
          name: myapp-vsvc
          routes:
            - primary
    steps:
      # 1. 设置权重
      - setWeight: 20
      # 2. Header 路由（带特定 Header 的请求全部路由到 canary）
      - setHeaderRoute:
          name: set-header-1
          match:
            - headerName: X-Canary
              headerValue:
                exact: "true"
      - pause: {}
      # 3. 流量镜像（将匹配流量镜像到 canary）
      - setMirrorRoute:
          name: mirror-route
          percentage: 35
          match:
            - method:
                exact: GET
              path:
                prefix: /api
      - pause: { duration: 10m }
      # 4. 清理路由
      - setHeaderRoute:
          name: set-header-1      # 仅 name = 删除路由
      - setMirrorRoute:
          name: mirror-route
      - setWeight: 100

apiVersion: argoproj.io/v1alpha1
kind: AnalysisTemplate
metadata:
  name: success-rate
spec:
  args:
    - name: service-name
    - name: stable-hash
    - name: latest-hash
  metrics:
    - name: success-rate
      interval: 5m                    # 每 5 分钟查询一次
      provider:
        prometheus:
          address: http://prometheus:9090
          query: |
            sum(rate(
              requests_total{service="{{args.service-name}}", status=~"2.."}[5m]
            )) / sum(rate(
              requests_total{service="{{args.service-name}}"}[5m]
            ))
      successCondition: result[0] >= 0.95    # 成功率 >= 95% 则成功
      failureLimit: 3                         # 连续 3 次失败则回滚
      consecutiveErrorLimit: 2              # 连续 2 次查询错误则暂停

# Background Analysis（canary 过程中持续监控）
strategy:
  canary:
    analysis:
      templates:
        - templateName: success-rate
      args:
        - name: service-name
          value: myapp-svc

# Inline Analysis（作为 canary step 执行）
steps:
  - setWeight: 20
  - analysis:
      templates:
        - templateName: smoke-tests
  - setWeight: 50

apiVersion: argoproj.io/v1alpha1
kind: Experiment
metadata:
  name: my-experiment
spec:
  duration: 1h                  # 实验持续 1 小时
  templates:
    - name: baseline
      specRef: stable           # 基于稳定版本 ReplicaSet 模板
    - name: canary
      specRef: canary           # 基于新版本 ReplicaSet 模板
      weight: 10                # 分配 10% 流量
  analyses:
    - name: compare
      templateName: mann-whitney    # 对比分析模板

steps:
  - setWeight: 20
  - experiment:
      duration: 1h
      templates:
        - name: baseline
          specRef: stable
        - name: canary
          specRef: canary
          weight: 10
      analyses:
        - name: mann-whitney
          templateName: mann-whitney

apiVersion: argoproj.io/v1alpha1
kind: Rollout
metadata:
  name: my-rollout
spec:
  replicas: 5                          # 期望副本数
  revisionHistoryLimit: 3              # 保留历史 ReplicaSet 数量（默认 10）
  minReadySeconds: 30                  # Pod 就绪后等待时间
  progressDeadlineSeconds: 600         # 更新超时时间（默认 600s）
  progressDeadlineAbort: false         # 超时是否自动中止
  rollbackWindow:
    revisions: 3                       # 快速回滚窗口（跳过 analysis）

  # Pod 模板（与 Deployment 一致）
  selector:
    matchLabels:
      app: myapp
  template:
    metadata:
      labels:
        app: myapp
    spec:
      containers:
        - name: app
          image: myapp:v1

  # 或使用 workloadRef 引用现有 Deployment
  workloadRef:
    apiVersion: apps/v1
    kind: Deployment
    name: my-deployment
    scaleDown: onsuccess              # never | onsuccess | progressively

  strategy:
    canary: { ... }                   # 或 blueGreen: { ... }

spec:
  workloadRef:
    apiVersion: apps/v1
    kind: Deployment
    name: existing-deployment
    scaleDown: progressively     # 逐步缩容原 Deployment

# Argo CD Application 管理 Rollout 定义
apiVersion: argoproj.io/v1alpha1
kind: Application
metadata:
  name: myapp
  namespace: argocd
spec:
  project: default
  source:
    repoURL: https://github.com/org/myapp-config
    path: kustomize/production
    targetRevision: main
  destination:
    server: https://kubernetes.default.svc
    namespace: myapp
  syncPolicy:
    automated:
      selfHeal: true
      prune: true

# 状态查看
kubectl argo rollouts get rollout <name>              # 查看 Rollout 状态
kubectl argo rollouts get rollout <name> --watch       # 实时观察
kubectl argo rollouts list rollouts                    # 列出所有 Rollout
kubectl argo rollouts history <name>                   # 查看历史版本

# 部署操作
kubectl argo rollouts set image <rollout> <container>=<image>  # 更新镜像
kubectl argo rollouts promote <name>                   # 手动推广
kubectl argo rollouts promote <name> --full            # 跳过所有步骤直接全量
kubectl argo rollouts abort <name>                     # 中止（回滚）
kubectl argo rollouts retry <name>                     # 重试已中止的 Rollout
kubectl argo rollouts restart <name>                   # 重启所有 Pod
kubectl argo rollouts pause <name>                     # 手动暂停
kubectl argo rollouts undo <name>                      # 回滚到上一版本

# 流量管理
kubectl argo rollouts set-weight <name> <weight>       # 手动设置流量权重
kubectl argo rollouts set-mirror <name> <percentage>   # 设置流量镜像

# Analysis
kubectl argo rollouts get analysisrun <name>           # 查看分析运行
kubectl argo rollouts terminate <name>                 # 终止分析运行

# Dashboard
kubectl argo rollouts dashboard                        # 启动 Web UI

# 版本
kubectl argo rollouts version                          # 查看 CLI 版本

# 1. 安装 Argo Rollouts Controller
# 2. 将 Deployment 改为 Rollout（apiVersion: argoproj.io/v1alpha1, kind: Rollout）
# 3. 添加 strategy 字段（canary 或 blueGreen）
# 4. kubectl apply 即可

# 或使用 workloadRef 渐进迁移（不修改现有 Deployment）