cadvisor 监控部署

#告警规则 参考：https://awesome-prometheus-alerts.grep.to/rules#docker-containers

Docker Root Dir: /home/docker

1.导入镜像

docker load < cadvisor.latest.tgz

2.docker-compose.yaml文件

cd /data/docker-swarm

cadvisor-docker-compose.yaml

version: "3.8"services: cadvisor: hostname: cadvisor #image: gcr.io/cadvisor/cadvisor:v0.47.1 image: google/cadvisor:latest networks: - monitor volumes: - /:/rootfs:ro - /var/run:/var/run:ro - /sys:/sys:ro - /data/docker:/var/lib/docker:ro - /dev/disk/:/dev/disk:ro ports: - target: 8080 published: 18080 protocol: tcp mode: host deploy: mode: global resources: limits: cpus: '0.50' memory: 300M command: - "--store_container_labels=false" - "--docker_only=true"networks: monitor: external: true

3.启停服务

docker stack rm cadvisor

docker stack deploy -c cadvisor-docker-compose.yaml cadvisor

4.prometheus配置

参考

# 进入prometheus配置目录cd /data/prometheus/conf# 配置需要监控的cadvisor实例连接地址（此处以staging环境为例）cat > targets//staging.yaml <<\EOF- labels: env: sz-prom-staging targets: - '192.168.102.213:18080' - '192.168.102.214:18080' - '192.168.102.235:18080' - '192.168.102.219:18080' - '192.168.102.236:18080' - '192.168.102.215:18080' - '192.168.102.220:18080' - '192.168.102.221:18080'EOF# 添加prometheus job配置cat >> prometheus.yml <<\EOF - job_name: cadvisor file_sd_configs: - files: - /etc/prometheus/targets/cadvisor/*.yamlEOF# 重载prometheus配置curl -vv --basic -u 'admin:meng@2022' -X POST http://192.168.10.19:9090/-/reload

参考：

/data/prometheus/conf/targets/node-exporter/

dev.yaml

- labels: cluster: sz-meng-cadvisor-dev targets: - '192.168.10.19:18080'#监控的label建议统一，只区分环境

#grafana如加载不了数据，需要查看面板的数据源是否准确