Prometheus监控系统部署实践

原创已于 2023-06-27 15:53:31 修改 · 502 阅读

0 ·

CC 4.0 BY-SA版权

文章标签：

#prometheus

于 2023-06-27 15:42:38 首次发布

本文详细介绍了如何通过Docker在内网服务器上部署Prometheus监控系统，包括Prometheusserver、Grafana、node-exporter、blackbox-exporter和alertmanager，以实现服务器资源监控、服务状态探测和异常警报功能。此外，还涉及到配置文件如prometheus.yml和alertmanager.yml的设置，以及数据持久化和启动脚本的编写。

用途：通过数据采集组件采集各服务器的内存，CPU，IO等使用情况，达到警戒线时邮件提醒用户；通过探针监控线上的服务是否正常，当服务不正常时邮件提醒用户。

说明：1 本文为实际操作说明，需要具备linux基础和docker基础知识。

2 本文部署在内网服务器，内网docker镜像仓库为x.x.x.x:5000;需要把相关镜像推送到私服，如服务器可以联网连上dockerhub ,镜像名可以去掉x.x.x.x:5000。

附官网prometheus架构图，方便理解：

本次部署采用docker部署，基本架构：

服务器	部署项目	用途
x.x.x.192:9090	prometheus server	用于收集和存储时间序列数据
x.x.x.192:3000	grafana	监控仪表盘，可视化监控数据
x.x.x.192:9115	blackbox-exporter	对HTTP.HTTPS,TCP,ICMP,DNZ的探测
x.x.x.192:9093	alertmanager	发出报警，常见的接收方式有：电子邮件，微信，钉钉, slack等。
x.x.x.104:9200等目标服务器	node-exporter	监控服务器CPU、内存、磁盘、I/O等信息
x.x.x.104:18080等目标服务器	cadvisor	Docker容器监控

1 prometheus server部署

#由于网关隔离，服务器端无法访问外网，先通过本地pull镜像后推送到内网镜像仓库。
docker pull prom/prometheus
docker tag prom/prometheus x.x.x.x:5000/prom/prometheus
docker push x.x.x.x:5000/prom/prometheus

#启动Prometheus服务，首次启动个空镜像出来，
docker run -it -d --name=prometheus x.x.x.x:5000/prom/prometheus

#将镜像内/prometheus copy到服务器，进行数据持久化
docker cp prometheus:/prometheus ./data/
docker cp prometheus:/etc/prometheus/prometheus.yml ./
chmod 777 -R data
chmod 777 prometheus.yml
#编写启动脚本
vi run.sh

#执行脚本
chmod 755 run.sh && ./run.sh

#!/bin/bash
DIR="$( cd "$( dirname "$0"  )" && pwd  )"
NAME="prometheus"
IMAGE="x.x.x.x:5000/prom/prometheus"
 
docker pull $IMAGE
docker rm -f $NAME
docker run -d -p 9090:9090 \
--restart=always \
-v $DIR/prometheus.yml:/etc/prometheus/prometheus.yml \
-v /etc/localtime:/etc/localtime \
-v $DIR/targets:/etc/prometheus/targets \
-v $DIR/data:/prometheus \
--name $NAME \
$IMAGE

目录截图

data目录存储数据

targets目录是配置文件：

prometheus.yml配置

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.
  
alerting:
  alertmanagers:
  - static_configs:
    - targets:
      - x.x.x.192:9093
 
#邮件监控
rule_files:
  - targets/rules/*.yml
  
scrape_configs:
  #资源监控
  - job_name: 'prometheus'
    file_sd_configs:
      - files:
        - targets/hosts_/*.yml
        refresh_interval: 5m
 
  #http监控
  - job_name: 'http_status'
    metrics_path: /probe
    params:
      module: [http_2xx]
    file_sd_configs:
      - files:
        - targets/http_2xx/*.yml
        refresh_interval: 5m
    relabel_configs:
      - source_labels: [__address__]
        target_label: __param_target
      - target_label: __address__
        replacement: x.x.x.192:9115
   
  # 服务器连通性监控
  - job_name: 'ping_status'
    metrics_path: /probe
    params:
      module: [icmp]
    file_sd_configs:
      - files:
        - targets/icmp/*.yml
        refresh_interval: 5m
    relabel_configs:
      - source_labels: [__address__]
        target_label: __param_target
      - target_label: __address__
        replacement: x.x.x.192:9115
  
  #端口监控
  - job_name: 'port_status'
    metrics_path: /probe
    params:
      module: [tcp_connect]
    file_sd_configs:
      - files:
        - targets/tcp_connect/*.yml
        refresh_interval: 5m
    relabel_configs:
      - source_labels: [__address__]
        target_label: __param_target
      - target_label: __address__
        replacement: x.x.x.192:9115

资源监控配置：

#配置资源监控
mkdir ./target/hosts_
vi ./target/hosts_/host.yml

target/hosts_/host.yml配置

- targets:
  - 'localhost:9090'
  - 'x.x.x.1:9200'
  - 'x.x.x.2:9200'
  - ...

端口监控配置:

#配置端口联通性
mkdir ./target/tcp_connect
vi ./target/tcp_connect/tcp_connect.yml

target/tcp_connect/tcp_connect.yml

- targets:
  - 'x.x.x.104:9200'
  labels:
    instance: 'x.x.x.104 node exporter'
    group: 'test'
- targets:
  - 'x.x.x.105:9200'
  labels:
    instance: 'x.x.x.105 node exporter'
    group: 'test'
...

http监控配置:

#http配置
mkdir ./target/htto_2xx
vi ./target/htto_2xx/http_2xx.yml

target/htto_2xx/http_2xx.yml

- targets:
  - 'http://xxx/jobqueue/swagger-ui.html'
  labels:
    instance: 'jobqueue'
    group: 'test'
- targets:
  - 'http://xxx/jobexecutor/swagger-ui.html'
  labels:
    instance: 'jobexecutor'
    group: 'test'
...

主机连通性配置

#http配置
mkdir ./target/icmp
vi ./target/icmp/icmp.yml

target/icmp/icmp.yml

- targets:
  - 'x.x.x.192'
  labels:
    instance: 'x.x.x.192'
    group: 'test'
- targets:
  - 'x.x.x.143'
  labels:
    instance: 'x.x.x.143'
    group: 'test'
...

预警规则配置

#http配置
mkdir ./target/rules
vi ./target/rules/node-up.yml

target/rules/node-up.yml

groups:
- name: Host
  rules:
  - alert: 主机状态报警
    expr: up == 0
    for: 1m
    labels:
      serverity: high
    annotations:
      summary: "{{$labels.instance}}:服务器宕机"
      description: "{{$labels.instance}}:服务器延时超过5分钟"
  - alert: CPU报警
    expr: 100 * (1 - avg(irate(node_cpu_seconds_total{mode="idle"}[2m])) by(instance)) > 90
    for: 1m
    labels:
      serverity: middle
    annotations:
      summary: "{{$labels.instance}}: High CPU Usage Detected"
      description: "{{$labels.instance}}: CPU usage is {{$value}}, above 90%"
 
  - alert: 内存报警
    expr: (node_memory_MemTotal_bytes - node_memory_MemAvailable_bytes) / node_memory_MemTotal_bytes * 100 > 90
    for: 1m
    labels:
      serverity: high
    annotations:
      summary: "{{$labels.instance}}: High Memory Usage Detected"
      description: "{{$labels.instance}}: Memory Usage i{{ $value }}, above 90%"
 
  - alert: 磁盘报警
    expr: 100 * (node_filesystem_size_bytes{fstype=~"xfs|ext4"} - node_filesystem_avail_bytes) / node_filesystem_size_bytes > 90
    for: 1m
    labels:
      serverity: middle
    annotations:
      summary: "{{$labels.instance}}: High Disk Usage Detected"
      description: "{{$labels.instance}}, mountpoint {{$labels.mountpoint}}: Disk Usage is {{ $value }}, above 90%"

  - alert: IO报警
    expr: 100-(avg(irate(node_disk_io_time_seconds_total[1m])) by(instance)* 100) < 60
    for: 1m
    labels:
      serverity: high
    annotations:
      summary: "{{$labels.mountpoint}} 流入磁盘IO使用率过高！"
      description: "{{$labels.mountpoint }} 流入磁盘IO大于60%(目前使用:{{$value}})"

  - alert: 网络报警
    expr: ((sum(rate (node_network_receive_bytes_total{device!~'tap.*|veth.*|br.*|docker.*|virbr*|lo*'}[5m])) by (instance)) / 100) > 102400
    for: 1m
    labels:
      serverity: high
    annotations:
    annotations:
      summary: "{{$labels.mountpoint}} TCP_ESTABLISHED过高！"
      description: "{{$labels.mountpoint }} TCP_ESTABLISHED大于1000%(目前使用:{{$value}}%)"

  - alert: 非200HTTP状态码
    expr: probe_http_status_code{job="http_status"} != 200
    for: 1m
    labels:
      serverity: high
    annotations:
      summary: "{{$labels.instance}} 非200HTTP状态码！"
      description: "{{$labels.mountpoint }} 非200HTTP状态码"

cadvisor

#cadvisor配置
mkdir ./target/cadvisor
vi ./target/cadvisor/cadvisor.yml

target/cadvisor/cadvisor.yml

- targets:
  - 'x.x.x.x:18080'

2 `Grafana` 部署

#由于网关隔离，服务器端无法访问外网，先通过本地pull镜像后推送到内网镜像仓库。

docker pull grafana/grafana

docker tag grafana/grafana x.x.x.x:5000/grafana/grafana

docker push x.x.x.x:5000/grafana/grafana

#启动grafana 服务，首次启动个空镜像出来，

docker run -it -d --name=grafana x.x.x.x/grafana/grafana

#将镜像内/var/lib/grafana copy到服务器，进行数据持久化

docker cp grafana:/var/lib/grafana ./data/

chmod 777 -R data

#编写启动脚本

vi run.sh

#执行脚本

chmod 755 run.sh && ./run.sh

run.sh脚本

#!/bin/bash
DIR="$( cd "$( dirname "$0"  )" && pwd  )"
NAME="grafana"
IMAGE="x.x.x.x:5000/grafana/grafana"
 
docker pull $IMAGE
docker rm -f $NAME
docker run -d -i -p 3000:3000 \
--name $NAME \
-v $DIR/data:/var/lib/grafana \
-v "/etc/localtime:/etc/localtime" \
-e "GF_SERVER_ROOT_URL=http://grafana.server.name" \
-e "GF_SECURITY_ADMIN_PASSWORD=admin" \
$IMAGE

3 node-exporter部署

docker run -d -p 9100:9100 \
-v "/proc:/host/proc" \
-v "/sys:/host/sys" \
-v "/:/rootfs" \
-v "/etc/localtime:/etc/localtime" \
--restart=always \
x.x.x.x:5000/prom/node-exporter \
--path.procfs /host/proc \
--path.sysfs /host/sys \
--collector.filesystem.ignored-mount-points "^/(sys|proc|dev|host|etc)($|/)"

4 blackbox-exporter部署

docker run -id \
--name blackbox-exporter \
--restart=always \
-p 9115:9115  \
x.x.x.x:5000/prom/blackbox-exporter

5 alertmanager 部署

alertmanager目录与run.sh在同一个目录

DIR="$( cd "$( dirname "$0"  )" && pwd  )"
docker rm -f alertmanager
docker run -d \
  -p 9093:9093 \
  --name alertmanager \
  --restart=always \
  -v $DIR/alertmanager:/etc/alertmanager \
  x.x.x.x:5000/prom/alertmanager

alertmanager目录中是alertmanager.yml文件及template目录;template目录下存email.tmpl文件

alertmanager.yml

# 全局配置项
global:
  resolve_timeout: 5m #超时,默认5min
  #邮箱smtp服务
  smtp_smarthost: 'mail.xxx.com.cn:25'
  smtp_from: 'xxx@xxx.com.cn'
  #smtp_auth_username: 'xxx@xxx.com.cn'
  #smtp_auth_password: 'xxx'
  #smtp_auth_password: ''
  #smtp_hello: ''
  smtp_require_tls: false

# 定义模板信息
templates:
  - 'template/*.tmpl'   # 路径

# 路由
route:
  group_by: ['alertname'] # 报警分组依据
  group_wait: 10s #组等待时间
  group_interval: 10s # 发送前等待时间
  repeat_interval: 4h #重复周期
  receiver: 'email' # 默认警报接收者

# 警报接收者
receivers:
- name: 'email' #警报名称
  email_configs:
  - to: 'xxx@qq.com,xxx@xxx.com.cn'  #接收警报的email
   # html: '{{ template "email.to.html" . }}' # 模板
    send_resolved: true


# 告警抑制
#inhibit_rules:
#  - source_match:
#      severity: 'critical'
#    target_match:
#      severity: 'warning'
#    equal: ['alertname', 'dev', 'instance']

email.tmpl

=========start==========<br>
告警程序: prometheus_alert <br>
告警级别: {{ .Labels.severity }} 级 <br>
告警类型: {{ .Labels.alertname }} <br>
故障主机: {{ .Labels.instance }} <br>
告警主题: {{ .Annotations.summary }} <br>
告警详情: {{ .Annotations.description }} <br>
触发时间: {{ .StartsAt.Format "2019-08-04 16:58:15" }} <br>
=========end==========<br>

grafana面板可以在https://grafana.com/grafana/dashboards查找模板，直接导入使用即可：