Kubernetes多副本Deployment中Dotnet Monitor正确监控方案咨询

你的单副本配置已经采用了Sidecar模式（每个Pod内运行dotnet monitor容器），这种模式天然支持多副本场景，只需要做以下调整和补充：

1. 调整Deployment副本数并修正路径问题 #

直接修改Deployment的replicas字段为你需要的数量，同时修正原配置中主容器诊断端口路径与挂载目录不匹配的问题（否则Sidecar无法连接到主容器的诊断socket）：

apiVersion: apps/v1
kind: Deployment 
namespace: test       
spec:
  progressDeadlineSeconds: 600
  replicas: 2  # 修改为目标副本数
  selector:
    matchLabels:
      app: manager-app  # 补充Selector标签，与Pod模板标签对齐
  template:  
    metadata:
      labels:
        app: manager-app  # 给Pod添加标签，用于后续Service和监控匹配
    spec:
      volumes:
      - name: diagnostics
        emptyDir: {}
      containers:
      - env:
        - name: DOTNETMONITOR_DiagnosticPort__ConnectionMode
          value: Listen
        - name: DOTNETMONITOR_DiagnosticPort__EndpointName
          value: /tmp/port.sock  # 修正路径，与挂载的/tmp目录对应
        - name: DOTNETMONITOR_Storage__DumpTempFolder
          value: /tmp/dumps  # 同步修正存储路径到挂载目录
        - name: DOTNETMONITOR_Urls
          value: http://localhost:52323           
        image: manager:master
        imagePullPolicy: Always
        name: chart
        ports:
        - containerPort: 10800
          protocol: TCP
        volumeMounts:
          - mountPath: /tmp
            name: diagnostics
      - name: sidecar
        image: dotnet/nightly/monitor
        ports:
        - containerPort: 52325
          name: metrics  # 给端口命名，方便后续Service引用
        args: ["collect","--urls", "http://*:52323", "--metricUrls", "http://*:52325", "--no-auth"]
        volumeMounts:
          - name: diagnostics
            mountPath: /tmp
        resources:
          limits:
            cpu: "1"
            memory: 500Mi
          requests:
            cpu: 200m
            memory: 500Mi

2. 创建Service暴露Metrics端点 #

创建ClusterIP类型的Service，匹配所有Pod的标签，将Sidecar的52325端口暴露给监控系统：

apiVersion: v1
kind: Service
metadata:
  name: manager-monitor-service
  namespace: test
spec:
  selector:
    app: manager-app  # 匹配Deployment Pod的标签
  ports:
  - name: metrics
    port: 52325
    targetPort: 52325  # 对应Sidecar的容器端口

3. 配置监控系统自动抓取Metrics #

以Prometheus为例，可通过PodMonitor或ServiceMonitor实现多副本Pod的Metrics自动发现：

示例PodMonitor配置（直接监控Pod） #

apiVersion: monitoring.coreos.com/v1
kind: PodMonitor
metadata:
  name: manager-pod-monitor
  namespace: test
spec:
  selector:
    matchLabels:
      app: manager-app  # 匹配Pod标签
  podMetricsEndpoints:
  - port: metrics  # 对应Sidecar容器的端口名
    path: /metrics
    interval: 15s

示例ServiceMonitor配置（通过Service代理监控） #

apiVersion: monitoring.coreos.com/v1
kind: ServiceMonitor
metadata:
  name: manager-service-monitor
  namespace: test
spec:
  selector:
    matchLabels:
      app: manager-app  # 需给Service添加对应标签
  endpoints:
  - port: metrics
    path: /metrics
    interval: 15s

关键注意事项 #

• 每个Pod的Sidecar仅监控当前Pod内的主容器，多副本间通过Pod级别的emptyDir隔离，不会互相干扰。
• 确保监控系统具备目标Namespace下的Pod/Service访问权限，RBAC配置需正确。

内容的提问来源于stack exchange，提问作者Iceforest