1、解压包

#server
wget  https://github.com/prometheus/prometheus/releases/download/v2.41.0/prometheus-2.41.0.linux-amd64.tar.gz
#alertmanager
wget https://github.com/prometheus/alertmanager/releases/download/v0.25.0/alertmanager-0.25.0.linux-amd64.tar.gz
#nodeExporter
wget https://github.com/prometheus/node_exporter/releases/download/v1.5.0/node_exporter-1.5.0.linux-amd64.tar.gz

2、安装prometheus-server

2.1、配置prometheus-server

# my global config
global:
  scrape_interval: 10s # Set the scrape interval to every 15 seconds. Default is every 1 minute.
  evaluation_interval: 10s # Evaluate rules every 15 seconds. The default is every 1 minute.  刷新监控规则

# Alertmanager的访问地址
alerting:
  alertmanagers:
    - static_configs:
        - targets: ['127.0.0.1:9093']

rule_files:
  - "/opt/prometheus/rules_conf/*.yml" 

scrape_configs:
  # prometheus server访问地址
  - job_name: "prometheus"
    static_configs:
      - targets: ["localhost:9090"]

  #自动发现有多种方式，这里根据文件配置自动发现主机
  - job_name: "node-exporter-discovery" 
    file_sd_configs:
      - refresh_interval: 1m
        files:
        - /opt/prometheus/node_conf/node_exporter.yaml

node_exporter.ymal配置文件说明

- targets:
  - 127.0.0.1:9100
  labels:
    idc: prd

image.png

2.3、启动prometheus-server

2.3.1、添加服务

 cat > /usr/lib/systemd/system/prometheus.service << 'EOF' 
[Unit]
Description=Prometheus
After=network.target

[Service]
User=prometheus
Group=prometheus
ExecStart=/opt/prometheus/prometheus --config.file=/opt/prometheus/prometheus.yml --web.enable-lifecycle --storage.tsdb.path=/data1/prometheus/data  --storage.tsdb.retention.time=7d --storage.tsdb.retention.size=10G --log.level=info



[Install]
WantedBy=multi-user.target
EOF

备注说明

--web.enable-lifecycle  #开启url刷新配置功能
--storage.tsdb.max-block-duration=2d #配置tsdb最大文件块时长2d
--storage.tsdb.min-block-duration=2h  #配置tsdb最小文件块时长
--storage.tsdb.retention=15d  # tsdb 保存的数据时长，默认90d

这一套参数解决了tsdb文件块mmap 内存不够的问题

2.3.2、更新启动服务

systemctl daemon-reload
systemctl enable prometheus.service --now
systemctl start prometheus.service

3、安装alertManager

3.1、配置alertManager

route:
  group_by: ['alertname']  #根据标签进行分组，分级告警。这里只是分组，没有分级，没有联动抑制告警
  group_wait: 10s    #分组的告警窗口
  group_interval: 10s  #
  repeat_interval: 2d  #重复间隔，生产10分钟即可
  receiver: 'web.hook'   #配置告警方式
receivers:
  - name: 'web.hook'
    webhook_configs:
      - url: 'http://127.0.0.1:5210/'   #告警方式配置

3.2、添加服务

cat > /usr/lib/systemd/system/alertmanager.service << 'EOF' 
[Unit]
Description=alertmanager
After=network.target

[Service]
User=prometheus
Group=prometheus
ExecStart=/opt/prometheus-alertmanager/alertmanager --config.file=/opt/prometheus-alertmanager/alertmanager.yml

[Install]
WantedBy=multi-user.target
EOF

3.3、更新启动服务

systemctl daemon-reload
systemctl enable alertmanager.service --now
systemctl start alertmanager.service

4、告警方式服务说明

需要python3环境，安装参考linux安装miniconda3
这里使用了一个简单的python 脚本，依赖说明： https://github.com/keijack/python-simple-http-server
安装说明：https://pypi.org/project/simple-http-server

脚本内容，主要功能是拦截alertManager的请求，将请请求转换成通知中心接受方式。alterManger不支持调用脚本的方式

from simple_http_server import route, server
from simple_http_server import Request
import json
import requests

@route("/", method=["GET", "POST", "PUT"])
def index(req=Request()):

    alert_url="http://xxxxxx.com/noticeSend"
    headers = {"Content-Type":"application/json"}
    request_data = json.loads(str(req.body,"utf-8"))
    alerts_data=request_data["alerts"]
    tmp_ip_arr = []
    for item in alerts_data:
        ip = item["labels"]["instance"].split(":")[0]
        tmp_ip_arr.append(ip)

    ip_str = ",".join(tmp_ip_arr)
    content = request_data["alerts"][0]["annotations"]["summary"]

    tmp_data={}
    tmp_data["userIds"]="123123"
    tmp_data["sms"]= {
        "templateParamList": [ip_str,content]
    }
    print(json.dumps(tmp_data))
    alert_req= requests.post(alert_url,json.dumps(tmp_data),headers={'Content-Type':'application/json'})
    if alert_req.status_code == 200:
        return {"msg": "alert success"}
    else:
        return {"msg": "alert failed"}


def main(*args):
    server.start(port=5210)

if __name__ == "__main__":
    main()

5、nodeExporter安装

5.1、解压nodeExport到指定位置

5.2、添加nodeExport服务

cat > /usr/lib/systemd/system/nodeExporter.service << 'EOF' 
[Unit]
Description=nodeExporter
After=network.target

[Service]
User=prometheus
Group=prometheus
ExecStart=/opt/node_exporter/node_exporter

[Install]
WantedBy=multi-user.target
EOF

5.3、更新启动服务

systemctl daemon-reload
systemctl enable nodeExporter.service --now
systemctl start nodeExporter.service

6、告警规则配置

6.1、修改告警规则

根据prometheus server的配置修改。这里用的general.yml配置文件

1673262087475.png

这个文档比较全https://awesome-prometheus-alerts.grep.to/rules.html#host-and-hardware

demo： 这是nodeExporter的主机告警规则，可以自行删减

groups:

- name: NodeExporter

  rules:

    - alert: HostOutOfMemory
      expr: 'node_memory_MemAvailable_bytes / node_memory_MemTotal_bytes * 100 < 10'
      for: 2m
      labels:
        severity: warning
      annotations:
        summary: Host out of memory
        description: "Node memory is filling up (< 10% left)\n  VALUE = {{ $value }}\n  LABELS = {{ $labels }}"



    - alert: HostUnusualNetworkThroughputIn
      expr: 'sum by (instance) (rate(node_network_receive_bytes_total[2m])) / 1024 / 1024 > 100'
      for: 5m
      labels:
        severity: warning
      annotations:
        summary: Hostnetworkthroughput 
        description: "Host network interfaces are probably receiving too much data (> 100 MB/s)\n  VALUE = {{ $value }}\n  LABELS = {{ $labels }}"

    - alert: HostUnusualNetworkThroughputOut
      expr: 'sum by (instance) (rate(node_network_transmit_bytes_total[2m])) / 1024 / 1024 > 100'
      for: 5m
      labels:
        severity: warning
      annotations:
        summary: unusual network throughput out
        description: "Host network interfaces are probably sending too much data (> 100 MB/s)\n  VALUE = {{ $value }}\n  LABELS = {{ $labels }}"

    - alert: HostUnusualDiskReadRate
      expr: 'sum by (instance) (rate(node_disk_read_bytes_total[2m])) / 1024 / 1024 > 50'
      for: 5m
      labels:
        severity: warning
      annotations:
        summary: unusual disk read rate
        description: "Disk is probably reading too much data (> 50 MB/s)\n  VALUE = {{ $value }}\n  LABELS = {{ $labels }}"

    - alert: HostUnusualDiskWriteRate
      expr: 'sum by (instance) (rate(node_disk_written_bytes_total[2m])) / 1024 / 1024 > 50'
      for: 2m
      labels:
        severity: warning
      annotations:
        summary: unusual disk write rate
        description: "Disk is probably writing too much data (> 50 MB/s)\n  VALUE = {{ $value }}\n  LABELS = {{ $labels }}"

    - alert: HostOutOfDiskSpace
      expr: '(node_filesystem_avail_bytes * 100) / node_filesystem_size_bytes < 10 and ON (instance, device, mountpoint) node_filesystem_readonly == 0'
      for: 2m
      labels:
        severity: warning
      annotations:
        summary: out of disk space
        description: "Disk is almost full (< 10% left)\n  VALUE = {{ $value }}\n  LABELS = {{ $labels }}"

    - alert: HostDiskWillFillIn24Hours
      expr: '(node_filesystem_avail_bytes * 100) / node_filesystem_size_bytes < 10 and ON (instance, device, mountpoint) predict_linear(node_filesystem_avail_bytes{fstype!~"tmpfs"}[1h], 24 * 3600) < 0 and ON (instance, device, mountpoint) node_filesystem_readonly == 0'
      for: 2m
      labels:
        severity: warning
      annotations:
        summary: disk will fill
        description: "Filesystem is predicted to run out of space within the next 24 hours at current write rate\n  VALUE = {{ $value }}\n  LABELS = {{ $labels }}"

    - alert: HostOutOfInodes
      expr: 'node_filesystem_files_free / node_filesystem_files * 100 < 10 and ON (instance, device, mountpoint) node_filesystem_readonly == 0'
      for: 2m
      labels:
        severity: warning
      annotations:
        summary: out of inodes
        description: "Disk is almost running out of available inodes (< 10% left)\n  VALUE = {{ $value }}\n  LABELS = {{ $labels }}"

    - alert: HostInodesWillFillIn24Hours
      expr: 'node_filesystem_files_free / node_filesystem_files * 100 < 10 and predict_linear(node_filesystem_files_free[1h], 24 * 3600) < 0 and ON (instance, device, mountpoint) node_filesystem_readonly == 0'
      for: 2m
      labels:
        severity: warning
      annotations:
        summary: inodes will fill
        description: "Filesystem is predicted to run out of inodes within the next 24 hours at current write rate\n  VALUE = {{ $value }}\n  LABELS = {{ $labels }}"

    - alert: HostUnusualDiskReadLatency
      expr: 'rate(node_disk_read_time_seconds_total[1m]) / rate(node_disk_reads_completed_total[1m]) > 0.1 and rate(node_disk_reads_completed_total[1m]) > 0'
      for: 2m
      labels:
        severity: warning
      annotations:
        summary: unusual disk read latency
        description: "Disk latency is growing (read operations > 100ms)\n  VALUE = {{ $value }}\n  LABELS = {{ $labels }}"

    - alert: HostUnusualDiskWriteLatency
      expr: 'rate(node_disk_write_time_seconds_total[1m]) / rate(node_disk_writes_completed_total[1m]) > 0.1 and rate(node_disk_writes_completed_total[1m]) > 0'
      for: 20m
      labels:
        severity: warning
      annotations:
        summary: unusual disk write latency
        description: "Disk latency is growing (write operations > 100ms)\n  VALUE = {{ $value }}\n  LABELS = {{ $labels }}"

    - alert: HostHighCpuLoad
      expr: '100 - (avg by(instance) (rate(node_cpu_seconds_total{mode="idle"}[2m])) * 100) > 80'
      for: 0m
      labels:
        severity: warning
      annotations:
        summary: high CPU load
        description: "CPU load is > 80%\n  VALUE = {{ $value }}\n  LABELS = {{ $labels }}"

    - alert: HostCpuIsUnderUtilized
      expr: '100 - (rate(node_cpu_seconds_total{mode="idle"}[30m]) * 100) < 20'
      for: 1w
      labels:
        severity: info
      annotations:
        summary: CPU is under utilized
        description: "CPU load is < 20% for 1 week. Consider reducing the number of CPUs.\n  VALUE = {{ $value }}\n  LABELS = {{ $labels }}"

    - alert: HostCpuStealNoisyNeighbor
      expr: 'avg by(instance) (rate(node_cpu_seconds_total{mode="steal"}[5m])) * 100 > 10'
      for: 0m
      labels:
        severity: warning
      annotations:
        summary: CPU steal noisy neighbor
        description: "CPU steal is > 10%. A noisy neighbor is killing VM performances or a spot instance may be out of credit.\n  VALUE = {{ $value }}\n  LABELS = {{ $labels }}"

    - alert: HostCpuHighIowait
      expr: 'avg by (instance) (rate(node_cpu_seconds_total{mode="iowait"}[5m])) * 100 > 5'
      for: 0m
      labels:
        severity: warning
      annotations:
        summary: Host CPU high iowait
        description: "CPU iowait > 5%. A high iowait means that you are disk or network bound.\n  VALUE = {{ $value }}\n  LABELS = {{ $labels }}"

    - alert: HostContextSwitching
      expr: '(rate(node_context_switches_total[5m])) / (count without(cpu, mode) (node_cpu_seconds_total{mode="idle"})) > 1000'
      for: 0m
      labels:
        severity: warning
      annotations:
        summary: Host context switching
        description: "Context switching is growing on node (> 1000 / s)\n  VALUE = {{ $value }}\n  LABELS = {{ $labels }}"

    - alert: HostSwapIsFillingUp
      expr: '(1 - (node_memory_SwapFree_bytes / node_memory_SwapTotal_bytes)) * 100 > 80'
      for: 2m
      labels:
        severity: warning
      annotations:
        summary: Host swap is filling up
        description: "Swap is filling up (>80%)\n  VALUE = {{ $value }}\n  LABELS = {{ $labels }}"

 
  


    - alert: HostRaidDiskFailure
      expr: 'node_md_disks{state="failed"} > 0'
      for: 2m
      labels:
        severity: warning
      annotations:
        summary: Host RAID disk failure
        description: "At least one device in RAID array on {{ $labels.instance }} failed. Array {{ $labels.md_device }} needs attention and possibly a disk swap\n  VALUE = {{ $value }}\n  LABELS = {{ $labels }}"


    - alert: HostOomKillDetected
      expr: 'increase(node_vmstat_oom_kill[1m]) > 0'
      for: 0m
      labels:
        severity: warning
      annotations:
        summary:  OOM kill detected
        description: "OOM kill detected\n  VALUE = {{ $value }}\n  LABELS = {{ $labels }}"

    - alert: HostEdacCorrectableErrorsDetected
      expr: 'increase(node_edac_correctable_errors_total[1m]) > 0'
      for: 0m
      labels:
        severity: info
      annotations:
        summary: Host EDAC Correctable Errors detected (instance {{ $labels.instance }})
        description: "Host {{ $labels.instance }} has had {{ printf \"%.0f\" $value }} correctable memory errors reported by EDAC in the last 5 minutes.\n  VALUE = {{ $value }}\n  LABELS = {{ $labels }}"



 
    - alert: HostRequiresReboot
      expr: 'node_reboot_required > 0'
      for: 4h
      labels:
        severity: info
      annotations:
        summary: Host requires reboot (instance {{ $labels.instance }})
        description: "{{ $labels.instance }} requires a reboot.\n  VALUE = {{ $value }}\n  LABELS = {{ $labels }}"

    - alert: node-exporter-down
      expr: up == 0 
      for: 1m
      labels: 
        severity: info
      annotations: 
        summary: " {{ $labels.instance }} 宕机了"  
        description: "instance: {{ $labels.instance }} \n- job: {{ $labels.job }} 关机了， 时间已经1分钟了。" 
        value: "{{ $value }}"
        instance: "{{ $labels.instance }}"

6.2、更新告警规则

这个有个前提是开启了 --web.enable-lifecycle 刷新更新参数，如果没有该参数必须要重启prometheus-server更新

curl -XPOST http://127.0.0.1:9090/-/reload

end，成功收到磁盘满的告警

附录： grafana安装https://www.jianshu.com/p/bd8293455c1d