部署prometheus、grafana、alertmanager

简介：由于资源有限，本实验用了两台机器

监控端：部署prometheus、grafana、alertmanager
被监控端：node_exporter、mysqld_exporter

一. 部署promethus

1. 下载

https://prometheus.io/download/

2. 解压

mkdir -p /data/prometheus

tar -zxvf /root/prometheus-2.42.0.linux-amd64.tar.gz -C /data/

cd /data

mv prometheus-2.42.0.linux-amd64/ prometheus

3. 部署

创建prometheus用户

useradd -s /sbin/nologin -M prometheus

mkdir -p /data/database/prometheus

chown -R prometheus:prometheus /data/database/prometheus/

配置systemctl启动项

vim /etc/systemd/system/prometheus.service

[Unit]
Description=Prometheus
Documentation=https://prometheus.io/
After=network.target
[Service]
Type=simple
User=prometheus
ExecStart=/data/prometheus/prometheus --web.enable-lifecycle --config.file=/data/prometheus/prometheus.yml --storage.tsdb.path=/data/database/prometheus
Restart=on-failure
[Install]
WantedBy=multi-user.target

4. 加载配置&启动服务

systemctl daemon-reload

systemctl start prometheus

systemctl status prometheus

systemctl enable prometheus

访问web页面，IP:9090
查看到监控的数据，IP:9090/metrics

二. 监控linux主机

1. 下载node_exporter

wget https://github.com/prometheus/node_exporter/releases/download/v1.5.0/node_exporter-1.5.0.linux-amd64.tar.gz

2.解压

tar -zxvf node_exporter-1.5.0.linux-amd64.tar.gz -C /data/

mv /data/node_exporter-1.5.0.linux-amd64/ /data/node_exporter

3. 配置systemctl启动项

vim /etc/systemd/system/node_exporter.service

[Unit]
Description=node_exporter
[Service]
ExecStart=/data/source.package/node_exporter-1.1.2.linux-amd64/node_exporter
ExecReload=/bin/kill -HUP $MAINPID
KillMode=process
Restart=on-failure
[Install]
WantedBy=multi-user.target

4. 加载配置&启动服务

systemctl daemon-reload

systemctl start node_exporter.service

systemctl status node_exporter.service

systemctl enable node_exporter.service

查看到被监控的数据，IP:9100/metrics

5. 监控端配置

在主配置文件最后加上下面三行

vim /data/prometheus/prometheus.yml

- job_name: 'agent1' #取一个job名称来代表被监控的机器
    static_configs:
    - targets: ['192.168.1.1:9100'] # 这里改成被监控机器的IP，后面端口接9100

测试prometheus.yaml文件有无报错

[root@VM-16-2-centos prometheus]# ./promtool check config prometheus.yml
Checking prometheus.yml
 SUCCESS: prometheus.yml is valid prometheus config file syntax

6. 重新加载prometheus配置文件

curl -X POST http://127.0.0.1:9090/-/reload，打开prometheus页面输入up查看是不是有对应的数据了
回到web管理界面 ——>点——>点Targets ——>可以看到多了一台监控目标

三. 监控mysql

1. 下载mysqld_exporter

wget https://github.com/prometheus/mysqld_exporter/releases/download/v0.14.0/mysqld_exporter-0.14.0.linux-amd64.tar.gz2

2. 解压

tar -zxvf mysqld_exporter-0.14.0.linux-amd64.tar.gz -C /data/

mv /data/mysqld_exporter-0.14.0.linux-amd64/ /data/mysqld_exporter

[root@VM-12-2-centos ~]# ls /data/mysqld_exporter/
LICENSE  mysqld_exporter  NOTICE

3. 安装mariadb数据库,并授权

yum -y install mariadb-server -y

systemctl start mariadb

[root@VM-12-2-centos ~]# mysql
Welcome to the MariaDB monitor.  Commands end with ; or g.
Your MariaDB connection id is 2
Server version: 5.5.68-MariaDB MariaDB Server

Copyright (c) 2000, 2018, Oracle, MariaDB Corporation Ab and others.

Type 'help;' or 'h' for help. Type 'c' to clear the current input statement.

MariaDB [(none)]>
MariaDB [(none)]> grant select,replication client,process ON *.* to 'mysql_monito'@'localhost' identified by '123';
Query OK, 0 rows affected (0.00 sec)

MariaDB [(none)]>
MariaDB [(none)]> flush privileges;
Query OK, 0 rows affected (0.00 sec)

MariaDB [(none)]>
MariaDB [(none)]> exit
Bye

4. 启动

nohup /usr/local/mysqld_exporter/mysqld_exporter --config.my-cnf=/usr/local/mysqld_exporter/.my.cnf &

5. 监控端配置

vim /data/prometheus/prometheus.yml

  - job_name: 'mysql' #取一个job名称来代表被监控的机器
    static_configs:
      - targets: ['192.168.1.1:9104'] # 这里改成被监控机器的IP，后面端口接9104

6. 重启prometheus

systemctl restart prometheus

回到web管理界面 ——>点——>点Targets ——>可以看到多了一台监控目标

四. 部署grafana

1. 下载

wget https://dl.grafana.com/enterprise/release/grafana-enterprise-9.3.6.linux-amd64.tar.gz

2. 解压

tar -zxvf grafana-enterprise-9.3.6.linux-amd64.tar.gz -C /data

mv grafana-9.3.6/ grafana

3. 修改初始化文件

备份

cp /data/grafana/conf/defaults.ini /data/grafana/conf/defaults.ini.bak

修改

vim /data/grafana/conf/defaults.ini

data = /data/database/grafana/data
logs = /data/database/grafana/log
plugins = /data/database/grafana/plugins
provisioning = /data/grafana/conf/provisioning/

4. 配置systemctl启动项

vim /etc/systemd/system/grafana-server.service

[Unit]
Description=Grafana
After=network.target
[Service]
User=grafana
Group=grafana
Type=notify
ExecStart=/data/grafana/bin/grafana-server -homepath /data/grafana/
Restart=on-failure
[Install]
WantedBy=multi-user.target

5. 加载配置&启动服务

systemctl daemon-reload

systemctl start grafana-server.service

systemctl status grafana-server.service

systemctl enable grafana-server.service

web页面：ip+3000
- 默认账号密码都是admin admin，登陆时需要修改密码。

6. 配置grafana

添加prometheus监控数据及模板，将grafana和prometheus关联起来，也就是在grafana中添加添加数据源
- 点击：设置->Data Source->Add data source->选择prometheus->url内填写http://IP:9090->save&test

点击：左边栏Dashboards“+”号内import->输入“8919”->load->更改name为“Prometheus Node”->victoriaMetrics选择刚创建的数据源“prometheus”
- 如要使用其他的模板，请到grafana的官网去查找 https://grafana.com/dashboards
设置完成后，点击”Dashboards”，->”victoriaMetrics”->”Prometheus Node”

五、部署alertmanager

1. 下载

https://prometheus.io/download/

2. 解压

tar -zxvf alertmanager-0.25.0.linux-amd64.tar.gz -C /data/

cd /data

mv alertmanager-0.25.0.linux-amd64/ alertmanager

chown -R prometheus:prometheus /data/alertmanager

mkdir -p /data/alertmanager/data

3. 配置报警系统altermanger服务

vim /data/alertmanager/alertmanager.yml（最初配置）

global:
  resolve_timeout: 5m
route:
  group_by: ['alertname']
  group_wait: 10s
  group_interval: 10s
  repeat_interval: 1h
  receiver: 'web.hook'
receivers:
- name: 'web.hook'
  webhook_configs:
  - url: 'http://127.0.0.1:5001/'
inhibit_rules:
  - source_match:
      severity: 'critical'
    target_match:
      severity: 'warning'
    equal: ['alertname', 'dev', 'instance']

4. 配置systemctl启动项

vim /etc/systemd/system/alertmanager.service

[Unit]
Description=Alertmanager
After=network.target
[Service]
Type=simple
User=prometheus
ExecStart=/data/alertmanager/alertmanager --config.file=/data/alertmanager/alertmanager.yml --storage.path=/data/alertmanager/data
Restart=on-failure
[Install]
WantedBy=multi-user.target

5. 加载配置&启动服务

systemctl daemon-reload

systemctl start alertmanager.service

systemctl status alertmanager.service

systemctl enable alertmanager.service

6. 配置promethues.yaml

备份

cp /data/prometheus/prometheus.yml /data/prometheus/prometheus.yml.bak

编辑

vim /data/prometheus/prometheus.yml （job_name中有几台监控的机器就写几行）

alerting:
  alertmanagers:
  - static_configs:
    - targets:
      - 192.168.1.1:9093

rule_files:
  - "/data/database/prometheus/rules/*.rules"

scrape_configs:
  # The job name is added as a label `job=` to any timeseries scraped from this config.
  - job_name: 'prometheus'

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

    static_configs:
    - targets: ['192.168.1.1:9090']


  - job_name: 'node'
    static_configs:
    - targets: ['192.168.1.2:9100']
    - targets: ['192.168.1.3:9100']
    - targets: ['192.168.1.4:9100']

测试prometheus.yaml文件有无报错（可以检测出rules文件有无报错）

cd /data/prometheus

./promtool check config prometheus.yml

[root@VM-16-2-centos prometheus]# ./promtool check config prometheus.yml
Checking prometheus.yml
  SUCCESS: 1 rule files found
 SUCCESS: prometheus.yml is valid prometheus config file syntax

Checking /data/database/prometheus/rules/node.rules
  SUCCESS: 21 rules found

7. 创建prometheus的规则文件

mkdir /data/database/prometheus/rules

vim /data/database/prometheus/rules/node.rules

groups:
  - name: Node-rules
    rules:
    - alert: Node-Down
      expr: up{job="node1"} == 0
      for: 1m
      labels:
        severity: 严重警告
        instance: "{{ $labels.instance }}"
      annotations:
        summary: "{{$labels.instance }} 节点已经宕机 1分钟"
        description: "节点宕机"

    - alert: Node-CpuHigh
      expr: (1 - avg by (instance) (irate(node_cpu_seconds_total{job="node",mode="idle"}[5m]))) * 100 > 80
      for: 1m
      labels:
        severity: 警告
        instance: "{{ $labels.instance }}"
      annotations:
        summary: "{{ $labels.instance }} cpu使用率超 80%"
        description: "CPU 使用率为 {{ $value }}%"

    - alert: Node-CpuIowaitHigh
      expr: avg by (instance) (irate(node_cpu_seconds_total{job="node",mode="iowait"}[5m])) * 100 > 80
      for: 1m
      labels:
        severity: 警告
        instance: "{{ $labels.instance }}"
      annotations:
        summary: "{{ $labels.instance }} CPU iowait 使用率超过 80%"
        description: "CPU iowait 使用率为 {{ $value }}%"

    - alert: Node-MemoryHigh
      expr: (1 - node_memory_MemAvailable_bytes{job="node"} / node_memory_MemTotal_bytes{job="node"}) * 100 > 80
      for: 1m
      labels:
        severity: 警告
        instance: "{{ $labels.instance }}"
      annotations:
        summary: "{{ $labels.instance }} Memory使用率超过 80%"
        description: "Memory 使用率为 {{ $value }}%"

    - alert: Node-Load5High
      expr: node_load5 > (count by (instance) (node_cpu_seconds_total{job="node",mode='system'})) * 1.2
      for: 1m
      labels:
        severity: 警告
        instance: "{{ $labels.instance }}"
      annotations:
        summary: "{{ $labels.instance }} Load(5m)过高,超出cpu核数1.2倍"
        description: "Load(5m)过高,超出cpu核数 1.2倍"

    - alert: Node-DiskRootHigh
      expr: (1 - node_filesystem_avail_bytes{job="node",fstype=~"ext.*|xfs",mountpoint ="/"} / node_filesystem_size_bytes{job="node",fstype=~"ext.*|xfs",mountpoint ="/"}) * 100 > 80
      for: 10m
      labels:
        severity: 警告
        instance: "{{ $labels.instance }}"
      annotations:
        summary: "{{ $labels.instance }} Disk(/ 分区) 使用率超过 80%"
        description: "Disk(/ 分区) 使用率为 {{ $value }}%"

    - alert: Node-DiskDataHigh
      expr: (1 - node_filesystem_avail_bytes{job="node",fstype=~"ext.*|xfs",mountpoint ="/data"} / node_filesystem_size_bytes{job="node",fstype=~"ext.*|xfs",mountpoint ="/data"}) * 100 > 80
      for: 10m
      labels:
        severity: 警告
        instance: "{{ $labels.instance }}"
      annotations:
        summary: "{{ $labels.instance }} Disk(/data 分区) 使用率超过 80%"
        description: "Disk(/data 分区) 使用率为 {{ $value }}%"

    - alert: Node-DiskReadHigh
      expr: irate(node_disk_read_bytes_total{job="node"}[5m]) > 20 * (1024 ^ 2)
      for: 1m
      labels:
        severity: 警告
        instance: "{{ $labels.instance }}"
      annotations:
        summary: "{{ $labels.instance }} Disk 读取字节数速率超过 20 MB/s"
        description: "Disk 读取字节数速率为 {{ $value }}MB/s"

    - alert: Node-DiskWriteHigh
      expr: irate(node_disk_written_bytes_total{job="node"}[5m]) > 20 * (1024 ^ 2)
      for: 1m
      labels:
        severity: 警告
        instance: "{{ $labels.instance }}"
      annotations:
        summary: "{{ $labels.instance }} Disk 写入字节数速率超过 20 MB/s"
        description: "Disk 写入字节数速率为 {{ $value }}MB/s"

    - alert: Node-DiskReadRateCountHigh
      expr: irate(node_disk_reads_completed_total{job="node"}[5m]) > 3000
      for: 1m
      labels:
        severity: 警告
        instance: "{{ $labels.instance }}"
      annotations:
        summary: "{{ $labels.instance }} Disk iops 每秒读取速率超过 3000 iops"
        description: "Disk iops 每秒读取速率为 {{ $value }}"

    - alert: Node-DiskWriteRateCountHigh
      expr: irate(node_disk_writes_completed_total{job="node"}[5m]) > 3000
      for: 1m
      labels:
        severity: 警告
        instance: "{{ $labels.instance }}"
      annotations:
        summary: "{{ $labels.instance }} Disk iops 每秒写入速率超过 3000 iops"
        description: "Disk iops 每秒写入速率为 {{ $value }}"

    - alert: Node-InodeRootUsedPercentHigh
      expr: (1 - node_filesystem_files_free{job="node",fstype=~"ext4|xfs",mountpoint="/"} / node_filesystem_files{job="node",fstype=~"ext4|xfs",mountpoint="/"}) * 100 > 80
      for: 10m
      labels:
        severity: 警告
        instance: "{{ $labels.instance }}"
      annotations:
        summary: "{{ $labels.instance }} Disk (/ 分区) inode 使用率超过 80%"
        description: "Disk (/ 分区) inode 使用率为 {{ $value }}%"

    - alert: Node-InodeBootUsedPercentHigh
      expr: (1 - node_filesystem_files_free{job="node",fstype=~"ext4|xfs",mountpoint="/data"} / node_filesystem_files{job="node",fstype=~"ext4|xfs",mountpoint="/data"}) * 100 > 80
      for: 10m
      labels:
        severity: 警告
        instance: "{{ $labels.instance }}"
      annotations:
        summary: "{{ $labels.instance }} Disk (/data 分区) inode 使用率超过 80%"
        description: "Disk (/data 分区) inode 使用率为 {{ $value }}%"

    - alert: Node-FilefdAllocatedPercentHigh
      expr: node_filefd_allocated{job="node"} / node_filefd_maximum{job="node"} * 100 > 80
      for: 10m
      labels:
        severity: 警告
        instance: "{{ $labels.instance }}"
      annotations:
        summary: "{{ $labels.instance }} Filefd 打开百分比超过 80%"
        description: "Filefd 打开百分比为 {{ $value }}%"

    - alert: Node-NetworkNetinBitRateHigh
      expr: avg by (instance) (irate(node_network_receive_bytes_total{device=~"eth0|eth1|ens33|ens37"}[1m]) * 8) > 20 * (1024 ^ 2) * 8
      for: 3m
      labels:
        severity: 警告
        instance: "{{ $labels.instance }}"
      annotations:
        summary: "{{ $labels.instance }} Network 接收比特数速率超过 20MB/s"
        description: "Network 接收比特数速率为 {{ $value }}MB/s"

    - alert: Node-NetworkNetoutBitRateHigh
      expr: avg by (instance) (irate(node_network_transmit_bytes_total{device=~"eth0|eth1|ens33|ens37"}[1m]) * 8) > 20 * (1024 ^ 2) * 8
      for: 3m
      labels:
        severity: 警告
        instance: "{{ $labels.instance }}"
      annotations:
        summary: "{{ $labels.instance }} Network 接收比特数速率超过 20MB/s"
        description: "Network 发送比特数速率为 {{ $value }}MB/s"

    - alert: Node-NetworkNetinPacketErrorRateHigh
      expr: avg by (instance) (irate(node_network_receive_errs_total{device=~"eth0|eth1|ens33|ens37"}[1m])) > 15
      for: 3m
      labels:
        severity: 警告
        instance: "{{ $labels.instance }}"
      annotations:
        summary: "{{ $labels.instance }} Network 接收错误包速率超过 15个/秒"
        description: "Network 接收错误包速率为 {{ $value }}个/秒"

    - alert: Node-NetworkNetoutPacketErrorRateHigh
      expr: avg by (instance) (irate(node_network_transmit_packets_total{device=~"eth0|eth1|ens33|ens37"}[1m])) > 15
      for: 3m
      labels:
        severity: 警告
        instance: "{{ $labels.instance }}"
      annotations:
        summary: "{{ $labels.instance }} Network 发送错误包速率超过 15个/秒"
        description: "Network 发送错误包速率为 {{ $value }}个/秒"

    - alert: Node-ProcessBlockedHigh
      expr: node_procs_blocked{job="node"} > 10
      for: 10m
      labels:
        severity: 警告
        instance: "{{ $labels.instance }}"
      annotations:
        summary: "{{ $labels.instance }} Process 当前被阻塞的任务的数量超过 10个"
        description: "Process 当前被阻塞的任务的数量为 {{ $value }}个"

    - alert: Node-TimeOffsetHigh
      expr: abs(node_timex_offset_seconds{job="node"}) > 3 * 60
      for: 2m
      labels:
        severity: 警告
        instance: "{{ $labels.instance }}"
      annotations:
        summary: "{{ $labels.instance }} 节点的时间偏差超过 3m"
        description: "节点的时间偏差为 {{ $value }}m"

    - alert: Node-TCPconnection
      expr: node_sockstat_TCP_tw{job="node"} > 15000
      for: 2m
      labels:
        severity: 警告
        instance: "{{ $labels.instance }}"
      annotations:
        summary: "{{ $labels.instance }} TCP 等待关闭的TCP连接数TIME_WAIT过高大于15000"
        description: "TCP 等待关闭的TCP连接数为 {{ $value }}"

8. 配置alertmanager邮件报警

vim /data/alertmanager/alertmanager.yml

# 全局配置项
global:
  resolve_timeout: 5m #处理超时时间，默认为5min
  smtp_smarthost: 'smtp.qq.com:465' #邮箱smtp服务器代理
  smtp_from: '111111112@qq.com' #发送邮箱名称
  smtp_auth_username: '111111112@qq.com' #邮箱名称
  smtp_auth_password: 'asdklfjwiehrqc' #邮箱授权码
  smtp_require_tls: false
  smtp_hello: 'qq.com'


# 定义报警模板
templates:
  - '/data/alertmanager/email.tmpl'

# 定义路由树信息
route:
  group_by: ['alertname'] #报警分组依据
  group_wait: 10s #最初即第一次等待多久时间发送一组警报的通知
  group_interval: 10s #在发送新警报前的等待时间
  repeat_interval: 10m #发送重复警报的周期 对于email配置中，此项不可以设置过低，否则将会由于邮件发送太多频繁，被smtp服务器拒绝
  receiver: 'email' #发送警报的接收者的名称，以下receivers name的名称

# 定义警报接收者信息
receivers:
  - name: 'email' # 警报
    email_configs: # 邮箱配置
    - to: '1111111112@qq.com, hello@163.com' #添加多个邮箱中间用,+空格分开
      html: '{{ template "email.html" . }}'
      send_resolved: true

# 一个inhibition规则是在与另一组匹配器匹配的警报存在的条件下，使匹配一组匹配器的警报失效的规则。两个警报必须具有一组相同的标签。
inhibit_rules:
  - source_match:
      severity: 'critical'
    target_match:
      severity: 'warning'
    equal: ['alertname', 'dev', 'instance']

9. 创建自定义报警模板

vim /data/alertmanager/email.tmpl

{{ define "email.html" }}
{{- if gt (len .Alerts.Firing) 0 -}}
{{- range $index, $alert := .Alerts -}}
 ======== 异常告警 ========
告警类型：{{ $alert.Labels.alertname }}
告警级别：{{ $alert.Labels.severity }}
告警实例：{{ $alert.Labels.instance }}
告警应用: {{ $alert.Labels.name }}
告警信息：{{ $alert.Annotations.summary }}
告警详情：{{ $alert.Annotations.description }}
告警时间：{{ $alert.StartsAt.Local }}
========== END ==========
 


{{- end }}

{{- end }}

{{- if gt (len .Alerts.Resolved) 0 -}}

{{- range $index, $alert := .Alerts -}}

 
======== 告警恢复 ========
告警类型：{{ $alert.Labels.alertname }}
告警级别：{{ $alert.Labels.severity }}
告警实例：{{ $alert.Labels.instance }}
告警详情：{{ $alert.Annotations.description }}
告警应用: {{ $alert.Labels.name }}
当前状态: OK
告警时间：{{ $alert.StartsAt.Local }}
恢复时间：{：{ $alert.EndsAt.Local }}
========== END ==========
 


{{- end }}

{{- end }}

{{- end }}

10. 重启服务

systemctl restart prometheus.service

systemctl restart alertmanager.service

11. 页面验证

web页面：ip+9090上点击alert选项查看是否存在规则

一. 部署promethus

1. 下载

2. 解压

3. 部署

4. 加载配置&启动服务

二. 监控linux主机

1. 下载node_exporter

2.解压

3. 配置systemctl启动项

4. 加载配置&启动服务

5. 监控端配置

6. 重新加载prometheus配置文件

三. 监控mysql

1. 下载mysqld_exporter

2. 解压

3. 安装mariadb数据库,并授权

4. 启动

5. 监控端配置

6. 重启prometheus

四. 部署grafana

1. 下载

2. 解压

3. 修改初始化文件

4. 配置systemctl启动项

5. 加载配置&启动服务

6. 配置grafana

五、部署alertmanager

1. 下载

2. 解压

3. 配置报警系统altermanger服务

4. 配置systemctl启动项

5. 加载配置&启动服务

6. 配置promethues.yaml

7. 创建prometheus的规则文件

9. 创建自定义报警模板

10. 重启服务

11. 页面验证

12. 邮件告警

服务器托管，北京服务器托管，服务器租用，机房机柜带宽租用