pmm docker-compose 运行试用

pmm 是一个很不错的mysql、mongo、postgresql 数据库监控平台，基于prometheus、grafana、exporter 算是一个比较完整
的数据库分析平台

参考架构图

环境准备

对于pg 以及mysql 的运行基于docker 官方镜像，添加了pmm client 包，同时为了添加指标，添加了gogs（git 服务），pgloader
mysql 数据迁移到pg，没有配置node exporter

• docker-compose文件

 

version: "3"

services: 

  pgloader-mysql:

    image: dimitri/pgloader

    command: pgloader mysql://root:dalongrong@mysql/gogs postgresql://postgres:dalong@postgres:5432/appdemo

  gogs:

    image: gogs/gogs

    ports:

      - "10022:22"

      - "3000:3000"

  pmmserver:

    image: percona/pmm-server:1.17

    ports: 

    - "80:80"

  mysqlexporter:

    image: prom/mysqld-exporter

    ports: 

    - "9104:9104"

    environment: 

    - "DATA_SOURCE_NAME=root:dalongrong@(mysql:3306)/"

  mysql:

    build: 

     context: ./mysql

     dockerfile: Dockerfile

    ports:

      - 3306:3306

    command: --character-set-server=utf8mb4 --collation-server=utf8mb4_unicode_ci

    environment:

      MYSQL_ROOT_PASSWORD: dalongrong

      MYSQL_DATABASE: gogs

      MYSQL_USER: gogs

      MYSQL_PASSWORD: dalongrong

      TZ: Asia/Shanghai

  postgres:

    build: 

     context: ./pg

     dockerfile: Dockerfile

    ports:

    - "5432:5432"

    environment:

    - "POSTGRES_PASSWORD:dalong"

  pgexporter:

    image: wrouesnel/postgres_exporter:latest

    ports:

    - 9187:9187

    environment:

    - DATA_SOURCE_NAME=postgresql://postgres:dalong@postgres:5432/postgres_exporter?sslmode=disable

• pg dockerfile

FROM postgres:latest

RUN set -ex; \

    apt-get update; \

 apt-get install -y gnupg apt-transport-https; \

 echo 'deb https://repo.percona.com/apt stretch main' > /etc/apt/sources.list.d/percona.list; \

 apt-get update && apt-get install -y --allow-unauthenticated pmm-client; \

 apt-get remove --purge -y gnupg apt-transport-https; \

 apt-get autoremove -y --purge; \

 rm -rf /var/lib/apt/lists/* \

 true

​

COPY init.sql /docker-entrypoint-initdb.d/

COPY initpmm.sh /root/

 

 

说明：
主要是initpmm.sh 文件,进行exporter metrics 注册

 

#!/bin/sh

while ! pmm-admin config --client-name pg1 --server pmmserver

do

    sleep 1

done

​

pmm-admin add external:metrics postgresql-demo pgexporter:9187

 

• mysql dockerfile

FROM mysql:5.7.16

RUN set -ex; \

    apt-get update; \

 apt-get install -y gnupg apt-transport-https; \

 echo 'deb https://repo.percona.com/apt stretch main' > /etc/apt/sources.list.d/percona.list; \

 apt-get update && apt-get install -y --allow-unauthenticated pmm-client; \

 apt-get remove --purge --force-yes -y apt-transport-https; \

 apt-get autoremove -y --force-yes --purge; \

 rm -rf /var/lib/apt/lists/* \

 true

COPY initpmm.sh /root/

 

说明：
主要是initpmm.sh 文件，进行exporter metrics 注册

 

#!/bin/sh

while ! pmm-admin config --client-name mysql1 --server pmmserver

do

    sleep 1

done

​

pmm-admin add external:metrics mysql-demo mysqlexporter:9104

 

 

启动&&配置

• 启动

docker-compose up -d

• 注册mysql 以及pg metrics

 docker-compose exec mysql sh /root/initpmm.sh

OK, PMM server is alive.

PMM Server | pmmserver 

Client Name | mysql1

Client Address | 172.24.0.8 

External metrics added.

​

docker-compose exec postgres sh /root/initpmm.sh

OK, PMM server is alive.

PMM Server | pmmserver 

Client Name | pg1

Client Address | 172.24.0.6 

External metrics added.

• prometheus target 效果

• grafana 效果

• 配置gogs导入部分数据

• pg 导入数据

docker-compose up pgloader-mysql

效果

pgloader-mysql_1 | WARNING:

pgloader-mysql_1 | Couldn't re-execute SBCL with proper personality flags (/proc isn't mounted? setuid?)

pgloader-mysql_1 | Trying to continue anyway.

pgloader-mysql_1 | 2019-08-05T14:26:59.075000Z LOG pgloader version "3.6.2~devel"

pgloader-mysql_1 | 2019-08-05T14:26:59.212000Z LOG Migrating from #<MYSQL-CONNECTION mysql://root@mysql:3306/gogs {1006496943}>

pgloader-mysql_1 | 2019-08-05T14:26:59.213000Z LOG Migrating into #<PGSQL-CONNECTION pgsql://postgres@postgres:5432/appdemo {10066DF6D3}>

pgloader-mysql_1 | 2019-08-05T14:27:01.364000Z LOG report summary reset

pgloader-mysql_1 | table name errors rows bytes total time

pgloader-mysql_1 | ----------------------------- --------- --------- --------- --------------

pgloader-mysql_1 | fetch meta data 0 119 0.256s

pgloader-mysql_1 | Create Schemas 0 0 0.013s

pgloader-mysql_1 | Create SQL Types 0 0 0.027s

pgloader-mysql_1 | Create tables 0 72 0.411s

pgloader-mysql_1 | Set Table OIDs 0 36 0.004s

pgloader-mysql_1 | ----------------------------- --------- --------- --------- --------------

pgloader-mysql_1 | gogs.access 0 0 0.032s

pgloader-mysql_1 | gogs.access_token 0 0 0.021s

pgloader-mysql_1 | gogs.action 0 0 0.078s

pgloader-mysql_1 | gogs.attachment 0 0 0.113s

pgloader-mysql_1 | gogs.collaboration 0 0 0.030s

pgloader-mysql_1 | gogs.deploy_key 0 0 0.068s

pgloader-mysql_1 | gogs.comment 0 0 0.122s

pgloader-mysql_1 | gogs.follow 0 0 0.075s

pgloader-mysql_1 | gogs.issue 0 0 0.102s

pgloader-mysql_1 | gogs.email_address 0 0 0.142s

pgloader-mysql_1 | gogs.hook_task 0 0 0.157s

pgloader-mysql_1 | gogs.issue_user 0 0 0.114s

pgloader-mysql_1 | gogs.issue_label 0 0 0.135s

pgloader-mysql_1 | gogs.login_source 0 0 0.130s

pgloader-mysql_1 | gogs.label 0 0 0.154s

pgloader-mysql_1 | gogs.mirror 0 0 0.154s

pgloader-mysql_1 | gogs.milestone 0 0 0.139s

pgloader-mysql_1 | gogs.org_user 0 0 0.122s

pgloader-mysql_1 | gogs.notice 0 0 0.154s

pgloader-mysql_1 | gogs.protect_branch 0 0 0.128s

pgloader-mysql_1 | gogs.protect_branch_whitelist 0 0 0.134s

pgloader-mysql_1 | gogs.public_key 0 0 0.141s

pgloader-mysql_1 | gogs.pull_request 0 0 0.147s

pgloader-mysql_1 | gogs.release 0 0 0.170s

pgloader-mysql_1 | gogs.repository 0 0 0.143s

pgloader-mysql_1 | gogs.star 0 0 0.146s

pgloader-mysql_1 | gogs.team 0 0 0.149s

pgloader-mysql_1 | gogs.team_user 0 0 0.149s

pgloader-mysql_1 | gogs.team_repo 0 0 0.166s

pgloader-mysql_1 | gogs.two_factor_recovery_code 0 0 0.151s

pgloader-mysql_1 | gogs.two_factor 0 0 0.155s

pgloader-mysql_1 | gogs."user" 0 1 0.3 kB 0.167s

pgloader-mysql_1 | gogs.upload 0 0 0.178s

pgloader-mysql_1 | gogs.watch 0 0 0.162s

pgloader-mysql_1 | gogs.version 0 1 0.0 kB 0.179s

pgloader-mysql_1 | gogs.webhook 0 0 0.167s

pgloader-mysql_1 | ----------------------------- --------- --------- --------- --------------

pgloader-mysql_1 | COPY Threads Completion 0 4 0.316s

pgloader-mysql_1 | Create Indexes 0 83 0.644s

pgloader-mysql_1 | Index Build Completion 0 83 0.222s

pgloader-mysql_1 | Reset Sequences 0 36 0.026s

pgloader-mysql_1 | Primary Keys 0 36 0.049s

pgloader-mysql_1 | Create Foreign Keys 0 0 0.000s

pgloader-mysql_1 | Create Triggers 0 0 0.001s

pgloader-mysql_1 | Set Search Path 0 1 0.002s

pgloader-mysql_1 | Install Comments 0 0 0.000s

pgloader-mysql_1 | ----------------------------- --------- --------- --------- --------------

pgloader-mysql_1 | Total import time ? 2 0.3 kB 1.260s

pmm-docker_pgloader-mysql_1 exited with code 0

• mysql 指标信息查看

• pg 指标查看

• 另外一种添加数据库以及metrics 的方法
  实际上pmm也提供了通过界面添加的方法，如下图

添加mysql

• 查询分析界面

说明

pmm 提供的指标模版还是很全的，包含了我们实际数据库分析中的很多方便的指标，可以帮助我们方便的分析系统的瓶颈，同时官方文档
也比较全的提供了操作手册（有pdf版本的）

参考资料

https://www.percona.com/doc/percona-monitoring-and-management/index.html
https://github.com/rongfengliang/pmm-learning