Clickhouse监控_使用Prometheus+Grafana来监控Clickhouse服务和性能指标

官方文档https://clickhouse.com/docs/zh/operations/monitoring

建议使用Prometheus+Grafana组合监控Clickhouse服务和性能指标

1、把prometheus的clickhouse_exporter下载到Clickhouse服务器,查找prometheus的clickhouse_exporter参见https://prometheus.io/docs/instrumenting/exporters/#databases,查看clickhouse_exporter的git信息可以登陆到https://github.com/ClickHouse/clickhouse_exporter上面查看

bash 复制代码
root@DAILACHDBUD002:~# git clone https://github.com/ClickHouse/clickhouse_exporter.git
Cloning into 'clickhouse_exporter'...
remote: Enumerating objects: 1185, done.
remote: Counting objects: 100% (306/306), done.
remote: Compressing objects: 100% (162/162), done.
remote: Total 1185 (delta 179), reused 164 (delta 138), pack-reused 879
Receiving objects: 100% (1185/1185), 1.36 MiB | 5.99 MiB/s, done.
Resolving deltas: 100% (359/359), done.

2、进入下载后的clickhouse_exporter目录,对其进行编译,由于https://github.com/ClickHouse/clickhouse_exporter上看到clickhouse_exporter相关信息是golang,所以安装是golang再编译

bash 复制代码
root@DAILACHDBUD002:~# cd clickhouse_exporter/
root@DAILACHDBUD002:~/clickhouse_exporter# ll clickhouse_exporter
ls: cannot access 'clickhouse_exporter': No such file or directory
root@DAILACHDBUD002:~/clickhouse_exporter# apt install golang-go
root@DAILACHDBUD002:~/clickhouse_exporter# go build clickhouse_exporter.go
root@DAILACHDBUD002:~/clickhouse_exporter# ll clickhouse_exporter
-rwxr-xr-x 1 root root 12591899 Dec 21 07:04 clickhouse_exporter*

3、编译好clickhouse_exporter后,配置clickhouse_exporter开机启动,其中/root/clickhouse_exporter/clickhouse_exporter -scrape_uri=http://default:W123@localhost:8123/ &此处必须加上后缀&,否则后面的systemctl start clickhouse_exporter会一直无法正常启然后systemctl status clickhouse_exporter会看到一直处于Active: activating (start)启动的状态而不是正常的Active: active (running)

bash 复制代码
root@DAILACHDBUD002:~# cat /root/script/clickhouse_exporter.sh
#/bin/sh
/root/clickhouse_exporter/clickhouse_exporter -scrape_uri=http://default:W123@localhost:8123/ &
bash 复制代码
root@DAILACHDBUD002:~# cat /usr/lib/systemd/system/clickhouse_exporter.service
[Unit]
Description=clickhouse_exporter
After=network.target

[Service]
User=root
Group=root
Type=forking
ExecStart=/bin/bash /root/script/clickhouse_exporter.sh
Restart=on-failure

[Install]
WantedBy=multi-user.target
bash 复制代码
root@DAILACHDBUD002:~# systemctl enable clickhouse_exporter
Created symlink /etc/systemd/system/multi-user.target.wants/clickhouse_exporter.service → /lib/systemd/system/clickhouse_exporter.service.

root@DAILACHDBUD002:~# systemctl start clickhouse_exporter

root@DAILACHDBUD002:~# systemctl status clickhouse_exporter
● clickhouse_exporter.service - clickhouse_exporter
     Loaded: loaded (/lib/systemd/system/clickhouse_exporter.service; enabled; vendor preset: enabled)
     Active: active (running) since Thu 2023-12-21 07:22:30 UTC; 3s ago

4、登陆Prometheus所在服务器,打开Prometheus配置文件vim /prometheus/prometheus-2.32.1/prometheus.yml,增加如下配置

bash 复制代码
 - job_name: "clickhouse_exporter"
    scrape_timeout: 10s
    metrics_path: /metrics
    static_configs:
      - targets: ['172.22.137.132:9116']
      - targets: ['172.22.137.133:9116']
      - targets: ['172.22.137.134:9116']
      - targets: ['172.22.137.136:9116']

5、在https://grafana.com/grafana/dashboards/中查找Clickhouse的模板,在Grafana监控配置页面导入Clickhouse的模板https://grafana.com/grafana/dashboards/882-clickhouse/或https://grafana.com/grafana/dashboards/14432-clickhouse-metrics-on-settings/,然后在Grafana监控配置页面的数据源选项中选择Prometheus

相关推荐
SelectDB技术团队1 天前
从 ClickHouse 迁移到 Doris:SQL 兼容、同步与验证清单
数据库·人工智能·sql·clickhouse·apache doris·selectdb·湖仓架构升级
川石课堂软件测试1 天前
涨薪技术|Prometheus之HTTP API中使用PromQL
网络协议·测试工具·jmeter·http·单元测试·postman·prometheus
java_logo3 天前
Docker 部署 ClickHouse Server:轻松搭建列式 OLAP 分析数据库平台
数据库·clickhouse·docker·私有化部署·olap·列式数据库·轩辕镜像
IT界的老黄牛3 天前
Prometheus TSDB 拆不出来、也存不了一年:4 条外部存储出路 + 容量测算
prometheus·时序数据库·监控·thanos·victoriametrics·remote_write
Ningcode_cloud3 天前
Kubernetes 弹性伸缩实验手册:从集群搭建到 HPA / VPA 自动扩缩容
云原生·k8s·prometheus
qq_452396234 天前
第十二篇:《数据采集:Grafana Alloy、Fluent Bit、Vector 的选型与配置》
java·贪心算法·grafana
2601_962097364 天前
可观测性实战:Prometheus+Grafana+OTel构建监控
grafana·prometheus·微服务架构·可观测性·opentelemetry
Joker-Full-stack4 天前
higress Grafana生产事故复盘:Loki 2.9 WAL 竞态 Bug 导致 5 天审计日志消失
bug·grafana·higress
IT界的老黄牛5 天前
Jenkins 监控一条龙:接入、看板 9964、11 条告警规则直接抄走
jenkins·grafana·prometheus·监控·ci-cd·告警规则
陈皮糖..5 天前
基于 Keepalived 的传统 Web 高可用架构的容器化改造与可观测性升级
运维·docker·性能优化·架构·云计算·prometheus