使用Prometheus+Grafana监控JVM

墨蓝 2023-06-23 13:30 32阅读 0赞

# 一、概述 #

## JMX Exporter ##

[https://github.com/prometheus/jmx\_exporter][https_github.com_prometheus_jmx_exporter]

它是Prometheus官方组件，作为一个JAVA Agent来提供本地JVM的metrics，并通过http暴露出来。这也是官方推荐的一种方式，可以获取进程的信息，比如CPU和内存使用情况。

Jmx\_exporter是以代理的形式收集目标应用的jmx指标，这样做的好处在于无需对目标应用做任何的改动。  
运行JMX exporter的方式：

java XXX -javaagent:/root/jmx_exporter/jmx_prometheus_javaagent-0.12.0.jar=3010:/root/jmx_exporter/config.yaml  -jar XXX.jar

## 下载 ##

目前最新版是0.12.0，下载链接为：

[https://repo1.maven.org/maven2/io/prometheus/jmx/jmx\_prometheus\_javaagent/0.12.0/jmx\_prometheus\_javaagent-0.12.0.jar][https_repo1.maven.org_maven2_io_prometheus_jmx_jmx_prometheus_javaagent_0.12.0_jmx_prometheus_javaagent-0.12.0.jar]

# 二、JMX Exporter配置 #

## simple-config.yml ##

我的prometheus安装路径为：/data/prometheus

新建配置文件simple-config.yml

mkdir /data/prometheus/jmx_exporter
    cd /data/prometheus/jmx_exporter
    wget https://repo1.maven.org/maven2/io/prometheus/jmx/jmx_prometheus_javaagent/0.12.0/jmx_prometheus_javaagent-0.12.0.jar
    vim simple-config.yml

内容如下：

---
    lowercaseOutputLabelNames: true
    lowercaseOutputName: true
    whitelistObjectNames: ["java.lang:type=OperatingSystem"]
    rules:
     - pattern: 'java.lang<type=OperatingSystem><>((?!process_cpu_time)\w+):'
       name: os_$1
       type: GAUGE
       attrNameSnakeCase: true

比如我有一个rms的java应用，启动方式为：

java -jar /data/rms/RMS.jar

使用JMX Exporter插件收集数据，需要改成这样：

java -javaagent:/data/prometheus/jmx_exporter/jmx_prometheus_javaagent-0.12.0.jar=3010:/data/prometheus/jmx_exporter/simple-config.yml -jar /data/rms/RMS.jar

注意：3010是代理端口，可以随意指定。

## prometheus.yml ##

修改配置文件，增加一个job\_name

vim /data/prometheus/prometheus.yml

最后一行增加

- job_name: 'java'
        scrape_interval: 30s
        static_configs:
        - targets: ['localhost:3010']

重启prometheus

/etc/init.d/prometheus-server restart

# 三、Grafana配置 #

## 导入Dashboard ##

模板链接为：

[https://grafana.com/grafana/dashboards/8563/revisions][https_grafana.com_grafana_dashboards_8563_revisions]

点击下面的Download，进行下载。

![1341090-20191219102932862-2060651565.png][]

登录Grafana，点击import

![1341090-20191219102906921-1700100510.png][]

上传json

![1341090-20191219103033737-474942423.png][]

数据源选择 Prometheus

![1341090-20191219103114994-302381244.png][]

最后打开刚刚导入的Dashboard，如下图

![1341090-20191219103305017-1726857864.png][]

本文参考链接：

[https://chanjarster.github.io/post/prom-grafana-jvm/][https_chanjarster.github.io_post_prom-grafana-jvm]

[https_github.com_prometheus_jmx_exporter]: https://github.com/prometheus/jmx_exporter
[https_repo1.maven.org_maven2_io_prometheus_jmx_jmx_prometheus_javaagent_0.12.0_jmx_prometheus_javaagent-0.12.0.jar]: https://repo1.maven.org/maven2/io/prometheus/jmx/jmx_prometheus_javaagent/0.12.0/jmx_prometheus_javaagent-0.12.0.jar
[https_grafana.com_grafana_dashboards_8563_revisions]: https://grafana.com/grafana/dashboards/8563/revisions
[1341090-20191219102932862-2060651565.png]: https://img2018.cnblogs.com/i-beta/1341090/201912/1341090-20191219102932862-2060651565.png
[1341090-20191219102906921-1700100510.png]: https://img2018.cnblogs.com/i-beta/1341090/201912/1341090-20191219102906921-1700100510.png
[1341090-20191219103033737-474942423.png]: https://img2018.cnblogs.com/i-beta/1341090/201912/1341090-20191219103033737-474942423.png
[1341090-20191219103114994-302381244.png]: https://img2018.cnblogs.com/i-beta/1341090/201912/1341090-20191219103114994-302381244.png
[1341090-20191219103305017-1726857864.png]: https://img2018.cnblogs.com/i-beta/1341090/201912/1341090-20191219103305017-1726857864.png
[https_chanjarster.github.io_post_prom-grafana-jvm]: https://chanjarster.github.io/post/prom-grafana-jvm/