第一次接触Prometheus还是2018年参加QCON北京会议时听的一场Prometheus的讲座，当时所在的公司监控系统主要使用的zabbix,应用级的监控使用的是Pinpoint的二次开发，所以对于Prometheus的了解还只是听闻，尽管很想使用，但是公司的zabbix已经非常成熟，包括后来小米的open-falcon也是没有机会去使用。直到今天我才想着要不测试环境的服务都用Prometheus吧，因现在的公司并没有监控系统，服务挂了以后都是使用的时候才会发现。我个人是比较喜欢docker的，于是有了此篇文章。

角色分配

  • Prometheus 采集数据
  • Grafana 用于图表展示
  • redis_exporter 用于收集redis的metrics
  • node-exporter 用于收集操作系统和硬件信息的metrics
  • cadvisor 用于收集docker的相关metrics

安装Docker

我通常使用如下命令安装最新版的docker

1
wget -O /etc/yum.repos.d/docker-ce.repo https://download.docker.com/linux/centos/docker-ce.repo && yum install -y docker-ce && systemctl enable docker.service && service docker start

安装Docker-Compose

可以使用如下命令安装最新版的docker-compose

1
2
curl -L https://github.com/docker/compose/releases/download/1.24.1/docker-compose-`uname -s`-`uname -m` -o /usr/local/bin/docker-compose
chmod +x /usr/local/bin/docker-compose

在安装过程中下载时间需要40分钟，仅仅15M的东西，所以可以通过如下方式手动下载：

1
echo "https://github.com/docker/compose/releases/download/1.24.1/docker-compose-`uname -s`-`uname -m`"

输出如下：

1
https://github.com/docker/compose/releases/download/1.24.1/docker-compose-Linux-x86_64

点击下载，完成后上传到服务器上，然后执行如下命令：

1
2
mv docker-compose-Linux-x86_64 /usr/local/bin/docker-compose
chmod +x /usr/local/bin/docker-compose

部署Prometheus和Grafana

新增Prometheus配置文件

首先，创建/data/prometheus/目录，然后创建prometheus.yml，填入如下内容：

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
global:
  scrape_interval:     15s # Set the scrape interval to every 15 seconds. Default is every 1 minute.
  evaluation_interval: 15s # Evaluate rules every 15 seconds. The default is every 1 minute.
  # scrape_timeout is set to the global default (10s).

# Alertmanager configuration
alerting:
  alertmanagers:
  - static_configs:
    - targets: ['10.10.170.161:9093']
      # - alertmanager:9093

# Load rules once and periodically evaluate them according to the global 'evaluation_interval'.
rule_files:
  - "node_down.yml"
  # - "first_rules.yml"
  # - "second_rules.yml"

# A scrape configuration containing exactly one endpoint to scrape:
# Here it's Prometheus itself.
scrape_configs:
  # The job name is added as a label `job=<job_name>` to any timeseries scraped from this config.
  - job_name: 'prometheus'

    # metrics_path defaults to '/metrics'
    # scheme defaults to 'http'.

    static_configs:
    - targets: ['10.10.170.161:9090']

  - job_name: 'redis'
    static_configs:
     - targets: ['10.10.170.161:9121']
       labels:
         instance: redis

  - job_name: 'node'
    scrape_interval: 8s
    static_configs:
     - targets: ['10.10.170.161:9100']
       labels:
         instance: node

  - job_name: 'cadvisor'
    static_configs:
     - targets: ['10.10.170.161:8088']
     labels:
         instance: cadvisor

接着创建node_down.yml,添加如下内容：

1
2
3
4
5
6
7
8
9
10
11
groups:
- name: node_down
  rules:
  - alert: InstanceDown
    expr: up == 0
    for: 1m
    labels:
      user: test
    annotations:
      summary: "Instance {{ $labels.instance }} down"
      description: "{{ $labels.instance }} of job {{ $labels.job }} has been down for more than 1 minutes."

创建docker-compose

继续在/data/prometheus/目录中创建docker-compose-prometheus.yml,添加如下内容：

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
version: '2'

networks:
    monitor:
        driver: bridge

services:
    prometheus:
        image: prom/prometheus
        container_name: prometheus
        hostname: prometheus
        restart: always
        volumes:
            - /data/prometheus/prometheus.yml:/etc/prometheus/prometheus.yml
            - /data/prometheus/node_down.yml:/etc/prometheus/node_down.yml
        ports:
            - "9090:9090"
        networks:
            - monitor

    grafana:
        image: grafana/grafana
        container_name: grafana
        hostname: grafana
        restart: always
        ports:
            - "3000:3000"
        networks:
            - monitor
    redis-exporter:
        image: oliver006/redis_exporter
        container_name: redis_exporter
        hostname: redis_exporter
        restart: always
        ports:
            - "9121:9121"
        networks:
            - monitor
        command:
            - '--redis.addr=redis://10.10.170.161:7000'
    node-exporter:
        image: quay.io/prometheus/node-exporter
        container_name: node-exporter
        hostname: node-exporter
        restart: always
        ports:
            - "9100:9100"
        networks:
            - monitor

    cadvisor:
        image: google/cadvisor:latest
        container_name: cadvisor
        hostname: cadvisor
        restart: always
        volumes:
            - /:/rootfs:ro
            - /var/run:/var/run:rw
            - /sys:/sys:ro
            - /var/lib/docker/:/var/lib/docker:ro
        ports:
            - "8088:8080"
        networks:
            - monitor

这里的端口号已经redis-addr请修改成自己的。

启动docker-compose

使用下面的命令启动docker-compose定义的容器

1
docker-compose -f /data/prometheus/docker-compose-prometheus.yml up -d

输入如下内容即代表启动成功：

1
2
3
4
5
6
Creating network "prometheus_monitor" with driver "bridge"
Creating cadvisor       ... done
Creating prometheus     ... done
Creating node-exporter  ... done
Creating redis_exporter ... done
Creating grafana        ... done

也可通过docker ps命令查看是否启动成功。如果要关闭并删除以上5个容器，只需要执行如下命令即可：

1
docker-compose -f /data/prometheus/docker-compose-monitor.yml down

同样也会输出如下日志：

1
2
3
4
5
6
7
8
9
10
11
Stopping cadvisor       ... done
Stopping node-exporter  ... done
Stopping grafana        ... done
Stopping redis_exporter ... done
Stopping prometheus     ... done
Removing cadvisor       ... done
Removing node-exporter  ... done
Removing grafana        ... done
Removing redis_exporter ... done
Removing prometheus     ... done
Removing network prometheus_monitor

打开http://10.10.170.161:9090/targets，如果State都是UP即代表Prometheus工作正常，如下图所示：
image-20210331153759922
当然，我第一次安装时只有一个是UP的，造成这个问题的主要原因是因为CentOS7的防火墙firewall导致的，将对应的端口添加到防火墙策略里即可：

1
2
3
4
5
6
firewall-cmd --zone=public --add-port=9100/tcp --permanent
firewall-cmd --zone=public --add-port=8088/tcp --permanent
firewall-cmd --zone=public --add-port=9121/tcp --permanent
firewall-cmd --zone=public --add-port=3000/tcp --permanent
firewall-cmd --zone=public --add-port=9090/tcp --permanent
firewall-cmd --reload

可通过如下命令查看端口策略是否已经生效

1
firewall-cmd --permanent --zone=public --list-ports

配置Grafana

打开http://10.10.170.161:3000，使用默认账号密码admin/admin登录并修改密码后，默认进来是创建数据库的页面，在如下图所示中，选择Prometheus。
image-20210331153853984
选择完成后，打开新的页面，在HTTP的URL中输入Prometheus的地址http://10.10.170.161:9090，并且将Access由Server更换为Browser，因为跨域的问题Server无法使用。点击保存并测试。
image-20210331153931767
现在数据源已经接通了，只剩下漂亮的报表了，在https://grafana.com/grafana/dashboards中搜索需要的Dashboard模版，并将其json文件下载下来。我本次主要监控docker与redis，于是只需要下载下面两个即可：

在Grafana菜单栏中第一个+号中，选择import
image-20210331154003850
将其上传后，选择prom为Prometheus也就是配置Prometheus是的Name的值，点击保存即可。等待一会儿就会出现如下界面：
image-20210331154050097