可靠性最佳实践¶

apiVersion: apps/v1
kind: Deployment
metadata:
  name: reliable-app
spec:
  replicas: 3
  selector:
    matchLabels:
      app: reliable-app
  template:
    metadata:
      labels:
        app: reliable-app
    spec:
      containers:
      - name: app
        image: myapp:v1.0
        ports:
        - containerPort: 8080

        # 存活探针 - 检测进程是否存活
        livenessProbe:
          httpGet:
            path: /healthz
            port: 8080
          initialDelaySeconds: 30    # 启动后 30 秒开始检测
          periodSeconds: 10           # 每 10 秒检测一次
          timeoutSeconds: 5           # 超时 5 秒
          failureThreshold: 3         # 连续失败 3 次重启
          successThreshold: 1         # 成功 1 次即恢复

        # 就绪探针 - 检测是否准备接收流量
        readinessProbe:
          httpGet:
            path: /ready
            port: 8080
          initialDelaySeconds: 5      # 启动后 5 秒开始检测
          periodSeconds: 5            # 每 5 秒检测一次
          timeoutSeconds: 3
          failureThreshold: 3
          successThreshold: 1

        # 启动探针 - 保护慢启动应用
        startupProbe:
          httpGet:
            path: /startup
            port: 8080
          initialDelaySeconds: 0
          periodSeconds: 10
          timeoutSeconds: 3
          failureThreshold: 30        # 最多等待 300 秒启动
          successThreshold: 1

        # 资源配置
        resources:
          requests:
            cpu: 200m
            memory: 256Mi
          limits:
            cpu: 1000m
            memory: 1Gi

apiVersion: v1
kind: ConfigMap
metadata:
  name: prometheus-rules
data:
  alerts.yaml: |
    groups:
    - name: pod-alerts
      interval: 30s
      rules:
      # Pod CPU 使用率告警
      - alert: PodHighCPU
        expr: |
          sum(rate(container_cpu_usage_seconds_total[5m])) by (pod, namespace) > 0.8
        for: 5m
        labels:
          severity: warning
        annotations:
          summary: "Pod CPU 使用率过高"
          description: "{{ $labels.namespace }}/{{ $labels.pod }} CPU 使用率超过 80%"

      # Pod 内存使用率告警
      - alert: PodHighMemory
        expr: |
          sum(container_memory_working_set_bytes) by (pod, namespace) / 
          sum(container_spec_memory_limit_bytes) by (pod, namespace) > 0.9
        for: 5m
        labels:
          severity: warning
        annotations:
          summary: "Pod 内存使用率过高"
          description: "{{ $labels.namespace }}/{{ $labels.pod }} 内存使用率超过 90%"

      # Pod 重启告警
      - alert: PodRestarting
        expr: |
          rate(kube_pod_container_status_restarts_total[15m]) > 0
        for: 5m
        labels:
          severity: critical
        annotations:
          summary: "Pod 频繁重启"
          description: "{{ $labels.namespace }}/{{ $labels.pod }} 在过去 15 分钟内重启"

      # Pod 状态异常
      - alert: PodNotReady
        expr: |
          kube_pod_status_phase{phase!="Running"} == 1
        for: 10m
        labels:
          severity: critical
        annotations:
          summary: "Pod 状态异常"
          description: "{{ $labels.namespace }}/{{ $labels.pod }} 状态为 {{ $labels.phase }}"