一、概述

本文介绍如何使用 Kube-Prometheus 监控 RocketMQ。通过部署 RocketMQ Exporter,将 RocketMQ 的消息队列指标暴露给 Prometheus 进行采集和监控。

核心组件:

  • RocketMQ Exporter:用于暴露 RocketMQ 指标的 Exporter
  • ServiceMonitor:Prometheus Operator 的 CRD,用于服务发现和指标采集配置

二、部署 RocketMQ Exporter

2.1 创建 Deployment

通过 Deployment 部署 RocketMQ Exporter,需要配置 NameServer 地址。

apiVersion: apps/v1
kind: Deployment
metadata:
  labels:
   app.kubernetes.io/component: exporter
   app.kubernetes.io/name: rocketmq-exporter
   app.kubernetes.io/part-of: kube-prometheus
  name: rocketmq-exporter
  namespace: monitoring
spec:
  progressDeadlineSeconds: 600
  replicas: 1
  revisionHistoryLimit: 10
  selector:
    matchLabels:
      app.kubernetes.io/component: exporter
      app.kubernetes.io/name: rocketmq-exporter
      app.kubernetes.io/part-of: kube-prometheus
  strategy:
    rollingUpdate:
      maxSurge: 25%
      maxUnavailable: 25%
    type: RollingUpdate
  template:
    metadata:
      labels:
        app.kubernetes.io/component: exporter
        app.kubernetes.io/name: rocketmq-exporter
        app.kubernetes.io/part-of: kube-prometheus
    spec:
      containers:
        - args:
            - '--rocketmq.config.namesrvAddr=172.16.15.121:9876'
          image: rocketmq-exporter:latest
          imagePullPolicy: Always
          name: rocketmq-exporter
          ports:
            - containerPort: 5557
              name: http
              protocol: TCP
          resources:
            limits:
              cpu: 500m
              memory: 1024Mi
            requests:
              cpu: 10m
              memory: 512Mi
          terminationMessagePath: /dev/termination-log
          terminationMessagePolicy: File
          volumeMounts:
            - mountPath: /home/exporter/logs
              name: exporter-log
            - mountPath: /etc/localtime
              name: date-config
      dnsPolicy: ClusterFirst
      imagePullSecrets:
        - name: default-secret-harbor
      restartPolicy: Always
      schedulerName: default-scheduler
      securityContext: {}
      terminationGracePeriodSeconds: 30
      volumes:
        - emptyDir: {}
          name: exporter-log
        - hostPath:
            path: /etc/localtime
            type: ''
          name: date-config

配置说明:

  • args:RocketMQ Exporter 启动参数
    • --rocketmq.config.namesrvAddr:RocketMQ NameServer 地址,多个地址用分号分隔
  • port: 5557:RocketMQ Exporter 默认暴露端口
  • imagePullSecrets:私有镜像仓库的认证信息
  • resources:资源限制,根据监控的 RocketMQ 规模调整

2.2 创建 Service

apiVersion: monitoring.coreos.com/v1
kind: ServiceMonitor
metadata:
  labels:
    app.kubernetes.io/component: exporter
    app.kubernetes.io/name: rocketmq-exporter
    app.kubernetes.io/part-of: kube-prometheus
  name: rocketmq-exporter
  namespace: monitoring
spec:
  endpoints:
    - interval: 15s
      port: http
      relabelings:
        - action: replace
          regex: (.*)
          replacement: $1
          sourceLabels:
            - __meta_kubernetes_pod_node_name
          targetLabel: instance
      scheme: http
  jobLabel: app.kubernetes.io/name
  targetLabels: [platform]
  selector:
    matchLabels:
      app.kubernetes.io/component: exporter
      app.kubernetes.io/name: rocketmq-exporter
      app.kubernetes.io/part-of: kube-prometheus

配置说明:

  • interval: 15s:每 15 秒采集一次指标
  • relabelings:指标重写规则,将 Pod 节点名作为 instance 标签值
  • targetLabels:将 Service 的 platform 标签添加到采集的指标中

三、验证部署

# 查看 Deployment 状态
kubectl get deployment rocketmq-exporter -n monitoring

# 查看 Pod 日志
kubectl logs -f <pod-name> -n monitoring

# 访问 Exporter 指标接口
kubectl port-forward svc/rocketmq-exporter 5557:5557 -n monitoring
curl http://localhost:5557/metrics

四、监控指标说明

RocketMQ Exporter 提供的核心监控指标:

  • rocketmq_broker_tps:Broker 的 TPS(每秒事务数)
  • rocketmq_broker_qps:Broker 的 QPS(每秒查询数)
  • rocketmq_producer_message_size:生产者消息大小
  • rocketmq_consumer_tps:消费者 TPS
  • rocketmq_consumer_message_size:消费者消息大小
  • rocketmq_group_get_latency:消费组消费延迟
  • rocketmq_group_diff:消费组积压消息数

五、注意事项

  1. NameServer 地址:需要配置正确的 RocketMQ NameServer 地址,支持多个地址
  2. 网络访问:确保 K8s 集群能够访问 RocketMQ NameServer
  3. 镜像准备:RocketMQ Exporter 需要提前构建并推送到镜像仓库
  4. 资源配置:根据 RocketMQ 集群规模调整 Exporter 的内存资源,避免 OOM

配置说明:

  • interval: 15s:每 15 秒采集一次指标
  • relabelings:指标重写规则,将 Pod 节点名作为 instance 标签值
  • targetLabels:将 Service 的 platform 标签添加到采集的指标中

三、验证部署

# 查看 Deployment 状态
kubectl get deployment rocketmq-exporter -n monitoring

# 查看 Pod 日志
kubectl logs -f <pod-name> -n monitoring

# 访问 Exporter 指标接口
kubectl port-forward svc/rocketmq-exporter 5557:5557 -n monitoring
curl http://localhost:5557/metrics

四、监控指标说明

RocketMQ Exporter 提供的核心监控指标:

  • rocketmq_broker_tps:Broker 的 TPS(每秒事务数)
  • rocketmq_broker_qps:Broker 的 QPS(每秒查询数)
  • rocketmq_producer_message_size:生产者消息大小
  • rocketmq_consumer_tps:消费者 TPS
  • rocketmq_consumer_message_size:消费者消息大小
  • rocketmq_group_get_latency:消费组消费延迟
  • rocketmq_group_diff:消费组积压消息数

五、注意事项

  1. NameServer 地址:需要配置正确的 RocketMQ NameServer 地址,支持多个地址
  2. 网络访问:确保 K8s 集群能够访问 RocketMQ NameServer
  3. 镜像准备:RocketMQ Exporter 需要提前构建并推送到镜像仓库
  4. 资源配置:根据 RocketMQ 集群规模调整 Exporter 的内存资源,避免 OOM