1. exporter介绍

exporter是prometheus监控中重要的组成部分,负责数据指标的采集。上篇文章介绍了prometheus server的相关内容,本文将介绍数据采集插件。官方给出的插件有node_exporter、blackbox_exporter、mysqld_exporter、snmp_exporter等,第三方的插件有redis_exporter,cadvisor等。
下面我将结合实际工作中的使用到的插件来分别介绍。

2. node_exporter

node_exporter主要用来采集机器的性能指标数据,包括cpu,内存,磁盘,io等基本信息。上边文章介绍promehteus server时已详细介绍了node_exporter,这里就不在赘述。

3. mysqld_exporter

mysqld_exporter主要用于监控采集mysql数据库服务器相关指标。
#下载二进制文件

root@db-100-51:~# cd /data/
root@db-100-51:/data# wget https://github.com/prometheus/mysqld_exporter/releases/download/v0.10.0/mysqld_exporter-0.10.0.linux-amd64.tar.gz
root@db-100-51:/data#  tar -xf mysqld_exporter-0.10.0.linux-amd64.tar.gz
root@db-100-51:/data# cd mysqld_exporter-0.10.0/
root@db-100-51:/data/mysqld_exporter-0.10.0# ls
LICENSE  mysqld_exporter  NOTICE
root@db-100-51:/data#  mkdir log

#mysqld_exporter需要连接到数据库,创建一个登录数据库的用户

mysql> GRANT PROCESS, REPLICATION CLIENT, SELECT ON *.* TO 'exporter'@'localhost' identified by cds8080808!';
mysql> flush privileges;

#创建用于连接数据库的配置文件

root@db-100-51:/data/mysqld_exporter-0.10.0# vim .my.cnf 
[client]
user=exporter
password=cds8080808!

#创建supervisor启动mysqld_exporter

root@db-100-51:/data/mysqld_exporter-0.10.0# vim /etc/supervisor/conf.d/mysqld_exporter.conf
[program:mysqld_exporter]
command = /data/mysqld_exporter-0.10.0/mysqld_exporter -config.my-cnf="/data/mysqld_exporter-0.10.0/.my.cnf"
autostart = true
autorestart = true
startsecs = 5
startretries = 3
redirect_stderr = true
stdout_logfile=/data/mysqld_exporter-0.10.0/log/out-mysqld_exporter.log
stderr_logfile=/data/mysqld_exporter-0.10.0/log/err-mysqld_exporter.log
stdout_logfile_maxbytes = 20MB
stdout_logfile_backups = 20

#启动mysqld_exporter

root@db-100-51:/data/mysqld_exporter-0.10.0# supervisorctl update mysqld_exporter
root@db-100-51:/data/mysqld_exporter-0.10.0# supervisorctl status mysqld_exporter

4. redis_exporter

redis_exporter主要用于监控采集redis数据库服务器相关指标。
#下载二进制文件

root@redis-node1-slave:~# cd /data/ wget https://github.com/oliver006/redis_exporter/releases/download/v0.24.0/redis_exporter-v0.24.0.linux-amd64.tar.gz
root@redis-node1-slave:/data# tar -xf redis_exporter-v0.24.0.linux-amd64.tar.gz
root@redis-node1-slave:/data# mkdir /data/redis_exporter-v0.24.0
root@redis-node1-slave:/data# mv redis_exporter redis_exporter-v0.24.0
root@redis-node1-slave:/data# cd redis_exporter-v0.24.0/
root@redis-node1-slave:/data/redis_exporter-v0.24.0# mkdir log

#配置supervisor启动redis_exporter

root@redis-node1-slave:/data/redis_exporter-v0.24.0# vim /etc/supervisor/conf.d/redis_exporter.conf
[program:redis_exporter]
command = /data/redis_exporter-v0.24.0/redis_exporter -redis.addr 10.13.0.100:6379 -redis.password cds-china   # 该出ip为redis服务器的ip

autostart = true
startsecs = 5
startretries = 3
redirect_stderr = true
stout_logfile = /data/redis_exporter-v0.24.0/log/out-redis_exporter.log
stderr_logfile = /data/redis_exporter-v0.24.0/log/err-redis_exporter.log
stdout_logfile_maxbytes = 20MB
stdout_logfile_backups = 20

#启动redis_exporter

root@redis-node1-slave:/data/redis_exporter-v0.24.0# supervisorctl update redis_exporter
root@redis-node1-slave:/data/redis_exporter-v0.24.0# supervisorctl status redis_exporter

5. black_exporter

black_exporter是prometheus社区提供的官方黑盒监控解决方案,其允许用户通过:http、https、dns、tcp以及icmp的方式对网络进行探测。我们利用icmp探针可以坚持网络是否通畅,利用http,https可以坚持网页是否可以正常访问,利用tcp检测服务端口判断服务是否正常。
#下载二进制文件

root@prometheus:~# cd /data/
root@prometheus:/data# wget https://github.com/prometheus/blackbox_exporter/releases/download/v0.13.0/blackbox_exporter-0.13.0.linux-amd64.tar.gz
root@prometheus:/data# tar -xf blackbox_exporter-0.13.0.linux-amd64.tar.gz 

#修改blackbox_exporter配置文件

root@prometheus:/data# cd blackbox_exporter-0.13.0/
root@prometheus:/data/blackbox_exporter-0.13.0# vim blackbox.yml 
modules:
  http_2xx:
    prober: http
    timeout: 5s
    http:
      preferred_ip_protocol: "ip4"
      no_follow_redirects: true
      valid_http_versions: ["HTTP/1.1", "HTTP/2"]
      valid_status_codes: [200,302]  # Defaults to 2xx
      method: GET
  http_post_2xx:
    prober: http
    http:
      method: POST
  tcp_connect:
    prober: tcp
    timeout: 5s
  icmp:
    prober: icmp
    timeout: 5s

#配置systemd启动blackbox_exporter

root@prometheus:/data/blackbox_exporter-0.13.0# vim /usr/lib/systemd/system/blackbox_exporter.service 

[Unit]
Description=Prometheus blackbox exporter
After=local-fs.target network-online.target network.target
Wants=local-fs.target network-online.target network.target

[Service]
User=root
Type=simple
WorkingDirectory=/data/blackbox_exporter-0.13.0/
ExecStart=/data/blackbox_exporter-0.13.0/blackbox_exporter --config.file=/data/blackbox_exporter-0.13.0/blackbox.yml
Restart=on-failure

[Install]
WantedBy=multi-user.target
root@prometheus:/data/blackbox_exporter-0.13.0# supervisorctl start blackbox_exporter 
root@prometheus:/data/blackbox_exporter-0.13.0# supervisorctl enable blackbox_exporter 
root@prometheus:/data/blackbox_exporter-0.13.0# supervisorctl status blackbox_exporter 

6. cadvisor

cadvisor是google开源的用于监控容器运行的工具。
#下载二进制文件

root@GZH-vSPC02:~# cd /data/
root@GZH-vSPC02:/data# wget https://github.com/google/cadvisor/releases/download/v0.33.0/cadvisor
root@GZH-vSPC02:/data# mkdir cadvisor-v0.33.0
root@GZH-vSPC02:/data# mv cadvisor cadvisor-v0.33.0/
root@GZH-vSPC02:/data# cd cadvisor-v0.33.0/
root@GZH-vSPC02:/data/cadvisor-v0.33.0# mkdir log

#配置supervisor启动cadvisor

root@GZH-vSPC02:/data/cadvisor-v0.33.0# vim /etc/supervisor/conf.d/cadvisor-server.conf 
[program:cadvisor-server]
command = /data/cadvisor-v0.33.0/cadvisor
autostart = true
autorestart = true
startsecs = 5
startretries = 3
redirect_stderr = true
stdout_logfile=/data/cadvisor-v0.33.0/log/out-cadvisor.log
stderr_logfile=/data/cadvisor-v0.33.0/log/err-cadvisor.log
stdout_logfile_maxbytes = 20MB
stdout_logfile_backups = 20
root@GZH-vSPC02:/data/cadvisor-v0.33.0# supervisorctl update cadvisor 
root@GZH-vSPC02:/data/cadvisor-v0.33.0# supervisorctl status cadvisor 

7. 配置prometheus服务端

上面布置了exporter,我们需要配置prometheus server来拿到exporter采集到数据。
#修改pormetheus配置文件,配置文件上篇文章已经介绍了,这里就不在解释

root@prometheus:~# cd /data/prometheus-2.4.3/
root@prometheus:/data/prometheus-2.4.3# vim prometheus.yml 
#my global config
global:
  scrape_interval:     30s # Set the scrape interval to every 15 seconds. Default is every 1 minute.
  evaluation_interval: 25s # Evaluate rules every 15 seconds. The default is every 1 minute.
  scrape_timeout: 25s #is set to the global default (10s).

#Alertmanager configuration
alerting:
  alertmanagers:
  - static_configs:
    - targets:
            - 10.13.0.80:9093     #此处ip为alertmanager的ip,有多个alertmanager时填写多个
      - 10.13.0.81:9093

#Load rules once and periodically evaluate them according to the global 'evaluation_interval'.
rule_files:
  - "/data/prometheus-2.4.3/rules/node_down.yml"
  - "/data/prometheus-2.4.3/rules/memory_over.yml"
  - "/data/prometheus-2.4.3/rules/disk_over.yml"
  - "/data/prometheus-2.4.3/rules/cpu_over.yml"
  - "/data/prometheus-2.4.3/rules/http_check.yml"
  - "/data/prometheus-2.4.3/rules/tcp_check.yml"
  - "/data/prometheus-2.4.3/rules/mysql_check.yml"
  - "/data/prometheus-2.4.3/rules/redis_down_check.yml"
  - "/data/prometheus-2.4.3/rules/redis_rule_check.yml"
  - "/data/prometheus-2.4.3/rules/container_down.yml"
  - "/data/prometheus-2.4.3/rules/ping_check.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=` to any timeseries scraped from this config.
  - job_name: 'prometheus'

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

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

  - job_name: 'GICHOST'
    file_sd_configs:
    - files: ['./node_exporter/host.json']

  - job_name: 'blackbox-http'
    scrape_interval: 5s
    metrics_path: /probe
    params:
      module: [http_2xx]  # Look for a HTTP 200 response.
    file_sd_configs:
    - files: ['./blackbox_exporter/http.json']
    relabel_configs:
      - source_labels: [__address__]
        target_label: __param_target
      - source_labels: [__param_target]
        target_label: instance
      - target_label: __address__
        replacement: 10.13.0.82:9115       #此处ip为blackbox服务的ip

  - job_name: 'blackbox-tcp'
    scrape_interval: 5s
    metrics_path: /probe
    params:
      module: [tcp_connect]
    file_sd_configs:
    - files: ['./blackbox_exporter/tcp.json']
    relabel_configs:
      - source_labels: [__address__]
        target_label: __param_target
      - source_labels: [__param_target]
        target_label: instance
      - target_label: __address__
        replacement: 10.13.0.82:9115

  - job_name: 'mysqld-exporter'
    file_sd_configs:
    - files: ['./mysqld_exporter/mysqld.json']

  - job_name: 'blackbox-ping'
    scrape_interval: 5s
    metrics_path: /probe
    params:
      module: [icmp]
    file_sd_configs:
    - files: ['./blackbox_exporter/ping.json']
    relabel_configs:
      - source_labels: [__address__]
        target_label: __param_target
      - source_labels: [__param_target]
        target_label: instance
      - target_label: __address__
        replacement: 10.13.0.82:9115

  - job_name: 'redis-exporter'
    file_sd_configs:
    - files: ['./redis_exporter/redis.json']

  - job_name: 'container-exporter'
    file_sd_configs:
    - files: ['./container_exporter/container.json']

8. 配置prometheus 主机监控文件和告警规则

8.1 主机监控文件

#mysqld_exporter主机监控文件

root@prometheus:/data/prometheus-2.4.3# mkdir mysqld_exporter
root@prometheus:/data/prometheus-2.4.3# cd mysqld_exporter/
root@prometheus:/data/prometheus-2.4.3/mysqld_exporter# vim mysqld.json
[
{
"targets":[
"10.13.0.10:9104"
],
"labels":{
"dbinstance":"db100.51"
}
}
]

#redis_exporter主机监控文件

root@prometheus:/data/prometheus-2.4.3# mkdir redis_exporter
root@prometheus:/data/prometheus-2.4.3# cd redis_exporter/
root@prometheus:/data/prometheus-2.4.3/redis_exporter# vim redis.json 
[
{
"targets":[
"10.13.0.11:9121",
"10.13.0.12:9121"
],
"labels":{
"service":"redis"
}
}
]

#blackbox_exporter主机监控文件

root@prometheus:/data/prometheus-2.4.3# mkdir blackbox_exporter
root@prometheus:/data/prometheus-2.4.3# cd blackbox_exporter
root@prometheus:/data/prometheus-2.4.3/blackbox_exporter# vim http.json
[
{
"targets":[
"http://10.13.0.13:6011/health",
"http://10.13.0.14:6011/health"
],
"labels":{
"service":"sre"
}
}
]
root@prometheus:/data/prometheus-2.4.3/blackbox_exporter# vim tcp.json
[
{
"targets":[
"10.13.0.14:13371",
"10.13.0.15:13371"
],
"labels":{
"service":"vspc"
}
}
]
root@prometheus:/data/prometheus-2.4.3/blackbox_exporter# vim ping.json
[
{
"targets":[
"10.13.0.16",
"10.13.0.17"
],
"labels":{
"service":"mysql"
}
}
]

#配置cadvisor主机监控文件

root@prometheus:/data/prometheus-2.4.3# mkdir container_exporter
root@prometheus:/data/prometheus-2.4.3# cd container_exporter/
root@prometheus:/data/prometheus-2.4.3/container_exporter# vim container.json
[
{
"targets":[
"10.13.0.18:8080",
"10.13.0.19:8080"
],
"labels":{
"service":"docker-monitor"
}
}
]

8.2 告警规则

node_exporter监控获取的cpu,磁盘,实例存活规则上篇文章已经介绍,这里不再介绍

root@prometheus:/data/prometheus-2.4.3# cd rules
root@prometheus:/data/prometheus-2.4.3/rules# ls
container_down.yml  disk_over.yml  mysql_check.yml  node_down.yml   redis_down_check.yml  tcp_check.yml
cpu_over.yml  http_check.yml  memory_over.yml  ping_check.yml  redis_rule_check.yml

#mysqld_exporter监控mysql数据告警规则

root@prometheus:/data/prometheus-2.4.3/rules# vim mysql_check.yml
groups:
- name: MySQLStatsAlert
  rules:
  - alert: MySQL is down
    expr: mysql_up == 0
    for: 1m
    labels:
      severity: critical
    annotations:
      summary: "Instance {{$labels.instance}} MySQL is down"
      description: "MySQL database is down. This requires immediate action!(current value is: {{$value}})"
  - alert: Mysql_High_QPS
    expr: rate(mysql_global_status_questions[5m]) > 8000 
    for: 1m
    labels:
      severity: warning
    annotations:
      summary: "{{$labels.instance}}: Mysql_High_QPS detected"
      description: "{{$labels.instance}}: Mysql opreation is more than 5000 per second ,(current value is: {{$value}})" 
  - alert: Mysql_Too_Many_Slow_Query
    expr: rate(mysql_global_status_slow_queries[30m]) > 3
    for: 1m
    labels:
      severity: warning
    annotations:
      summary: "{{$labels.instance}}: Mysql_Too_Many_Slow_Query detected"
      description: "{{$labels.instance}}: Mysql current Slow_Query Sql is more than 3 ,(current value is: {{$value}})"
  - alert: Mysql_Deadlock
    expr: mysql_global_status_innodb_deadlocks > 300
    for: 1m
    labels:
      severity: warning
    annotations:
      summary: "{{$labels.instance}}: Mysql_Deadlock detected"
      description: "{{$labels.instance}}: Mysql Deadlock was found ,(current value is: {{$value}})"
  - alert: open files high
    expr: mysql_global_status_innodb_num_open_files > (mysql_global_variables_open_files_limit) * 0.75
    for: 1m
    labels:
      severity: warning
    annotations:
      summary: "Instance {{$labels.instance}} open files high"
      description: "Open files is high. Please consider increasing open_files_limit.(current value is: {{$value}})"
  - alert: Used more than 80% of max connections limited 
    expr: mysql_global_status_max_used_connections > mysql_global_variables_max_connections * 0.8
    for: 1m
    labels:
      severity: warning
    annotations:
      summary: "Instance {{$labels.instance}} Used more than 80% of max connections limited"
      description: "Used more than 80% of max connections limited.(current value is: {{$value}})"
  - alert: InnoDB Log File size is too small
    expr: mysql_global_variables_innodb_log_file_size < 16777216 
    for: 1m
    labels:
      severity: warning
    annotations:
      summary: "Instance {{$labels.instance}} InnoDB Log File size is too small"
      description: "The InnoDB Log File size is possibly too small. Choosing a small InnoDB Log File size can have significant performance impacts.(current value is: {{$value}})"
  - alert: Binary Log is disabled
    expr: mysql_global_variables_log_bin != 1
    for: 1m
    labels:
      severity: warning
    annotations:
      summary: "Instance {{$labels.instance}} Binary Log is disabled"
      description: "Binary Log is disabled. This prohibits you to do Point in Time Recovery (PiTR).(current value is: {{$value}})"
  - alert: IO thread stopped
    expr: mysql_slave_status_slave_io_running != 1
    for: 1m
    labels:
      severity: critical
    annotations:
      summary: "Instance {{$labels.instance}} IO thread stopped"
      description: "IO thread has stopped. This is usually because it cannot connect to the Master any more.(current value is: {{$value}})"
  - alert: SQL thread stopped
    expr: mysql_slave_status_slave_sql_running != 1
    for: 1m
    labels:
      severity: critical
    annotations:
      summary: "Instance {{$labels.instance}} Sync Binlog is enabled"
      description: "SQL thread has stopped. This is usually because it cannot apply a SQL statement received from the master.(current value is: {{$value}})"
  - alert: Slave lagging behind Master
    expr: rate(mysql_slave_status_seconds_behind_master[1m]) >30 
    for: 1m
    labels:
      severity: warning 
    annotations:
      summary: "Instance {{$labels.instance}} Slave lagging behind Master"
      description: "Slave is lagging behind Master. Please check if Slave threads are running and if there are some performance issues!(current value is: {{$value}})"

#redis_exporter监控redis指标告警规则

root@prometheus:/data/prometheus-2.4.3/rules# vim redis_down_check.yml
groups:
- name: redis检测规则
  rules:
  - alert: redis存活检测
    expr: redis_up{job="redis-exporter"} == 0
    for: 1m
    annotations:
      description: "机器:{{ $labels.instance }} 所属 job:{{ $labels.job }} redis宕机,请检查!"
      summary: "redis服务"
root@prometheus:/data/prometheus-2.4.3/rules# vim redis_rule_check.yml
    groups:
- name: RedisStatsAlert
  rules:
  - alert: last create rdb failed                       
    expr: redis_rdb_last_bgsave_status != 1
    for: 1m
    labels:
      severity: warning
    annotations:
      summary: " Instance {{ $labels.instance }} rdb_last_bgsave_status  "
      description: "last create rdb failed"
  - alert: Redis linked too many clients
    expr: redis_connected_clients / redis_config_maxclients * 100 > 80
    for: 1m
    labels:
      severity: warning
    annotations:
      summary: "Instance {{ $labels.instance }} Redis linked clients too many"
      description: "Redis linked clients too many. This requires immediate action!"
  - alert: master link status failed                                     
    expr: redis_master_link_up == 0
    for: 1m
    labels:
      severity: warning
    annotations:
      summary: "Instance {{ $labels.instance }} link failed"
      description: "redis_master_link=0 link failed"
  - alert: last AOF failed
    expr: redis_aof_last_bgrewrite_status != 1
    for: 1m
    labels:
      severity: warning
    annotations:
      summary: "Instance {{ $labels.instance }} redis aof last rewrite duration sec"
      description: "last AOF failed"
  - alert: Redis Cluster State Wrong
    expr: redis_cluster_state != 1
    for: 1m
    labels:
      severity: warning
    annotations:
      summary: "Instance {{ $labels.instance }} redis cluster status wrong"
      description: "Redis Cluster State Wrong"

#blackbox_exporter监控指标告警规则

root@prometheus:/data/prometheus-2.4.3/rules# vim ping_check.yml
groups:
- name: 机器网络存活检测
  rules:
  - alert: 网络检测
    expr: probe_success{job="blackbox-ping"} == 0
    for: 1m
    annotations:
      description: "机器:{{ $labels.instance }} 所属 job:{{ $labels.job }} 网络不通或者宕机超过1分钟,请检查!"
      summary: "网络检测"
root@prometheus:/data/prometheus-2.4.3/rules# vim http_check.yml
groups:
- name: 服务检测规则
  rules:
  - alert: http服务检测
    expr: probe_success{job="blackbox-http"} == 0
    for: 1m
    annotations:
      description: "机器:{{ $labels.instance }} 所属 job:{{ $labels.job }} http状态码: {{ printf `probe_http_status_code{instance='%s'}` $labels.instance | query | first | value }} http检测失败,请检查!"
      summary: "http检测"
root@prometheus:/data/prometheus-2.4.3/rules# vim tcp_check.yml
groups:
- name: 服务检测规则
  rules:
  - alert: tcp服务检测
    expr: probe_success{job="blackbox-tcp"} == 0
    for: 1m
    annotations:
      description: "机器:{{ $labels.instance }} 所属 job:{{ $labels.job }} tcp检测失败,请检查!"
      summary: "tcp检测"

#cadvisor监控指标告警规则

root@prometheus:/data/prometheus-2.4.3/rules# vim container_down.yml 

groups:
- name: 容器存活报警规则
  rules:
  - alert: DockerInstanceDown
    expr: absent(container_last_seen{name="catcti"}) == 1
    for: 1m
    annotations:
      description: "cacti容器:{{ $labels.name }} (所属主机{{ $labels.instance }}) 已经异常退出超过1分钟,请检查!"
      summary: "容器:Instance {{ $labels.name }} 存活检测"

参考文档:

https://github.com/prometheus
https://github.com/oliver006/redis_exporter
https://github.com/google/cadvisor