grafana phlare 试用

主要是简单试用下grafana phlare

环境准备

• docker-compose 文件

version: '3'

services:

  phlare:

    image: grafana/phlare:latest

    ports:

      - "4100:4100"

    command: -config.file=/etc/phlare/config.yaml

    volumes:

      - ./phlare.yaml:/etc/phlare/config.yaml

      - data:/data

    networks:

      - phlare

 

  grafana:

    image: grafana/grafana:main

    ports:

      - "3000:3000"

    environment:

      - GF_FEATURE_TOGGLES_ENABLE=flameGraph

      - GF_AUTH_ANONYMOUS_ENABLED=true

      - GF_AUTH_ANONYMOUS_ORG_ROLE=Admin

      - GF_DIAGNOSTICS_PROFILING_ENABLED=true

      - GF_DIAGNOSTICS_PROFILING_ADDR=0.0.0.0

      - GF_DIAGNOSTICS_PROFILING_PORT=6060

    volumes:

      - ./datasource.yaml:/etc/grafana/provisioning/datasources/datasources.yml

    networks:

      - phlare

volumes:

  data:

networks:

  phlare:

• 配置文件
  datasource.yaml

 

apiVersion: 1

datasources:

  - access: proxy

    basicAuth: false

    editable: true

    isDefault: true

    name: phlare

    type: phlare

    uid: phlare

    url: http://phlare:4100

    version: 1

phlare.yaml

scrape_configs:

  - job_name: "phlare"

    scrape_interval: "15s"

    static_configs:

      - targets: ["phlare:4100"]

  - job_name: "grafana"

    scrape_interval: "15s"

    static_configs:

      - targets: ["grafana:6060"]

  - job_name: "cp-java-springboot"

    scrape_interval: "15s"

    static_configs:

      - targets: ["docker.for.mac.host.internal:9999"]

    profiling_config:

      pprof_config:

        block: { enabled: false }

        goroutine: { enabled: false }

        memory: { enabled: false }

        mutex: { enabled: false }

• spring boot 项目
  下载地址: https://github.com/rongfengliang/jpprof-springboot-learning/releases/download/1.0/jpprofappv2-0.0.1-SNAPSHOT.jar
  运行需要试用jdk 11

 

java -jar jpprofappv2-0.0.1-SNAPSHOT.jar

运行&&效果

• 启动
  需要springboot 应用以及docker-compose 服务

 

docker-compose up -d

java -jar jpprofappv2-0.0.1-SNAPSHOT.jar

• 压测

ab -n 20000 -c 10 http://127.0.0.1:9999/demo

• 效果

 

 

说明

grafana phlare 目前来说使用还是比较简单的，核心是我们的代码需要调整下，添加支持profile 的地址，使用上与prometheus 服务发现的机制是类似的，体验上还是不错的,配置管理上与parca 是类似的套路，后期可以关注下此项目的迭代，目前来说与grafana 比较好的集成是一个很大的亮点，同时系统的扩展性也是不错的（技术栈上复用了不少grafana 其他工具的公共能力）

参考资料

https://github.com/grafana/phlare/tree/main/tools/docker-compose
https://grafana.com/docs/phlare/latest/operators-guide/get-started/
https://github.com/parca-dev/parca
https://github.com/rongfengliang/jpprof-springboot-learning