大数据组件监控体系构建

大数据监控体系构建

基于开源xxx_exporter+promethues+grafana的构建监控系统

1.概况

其中，exporter一般是使用来采集各种组件运行时的指标数据；
promethues构建指标时序数据库；
grafana构建指标显示面板；

目前已有各种docker容器方便的构建各种监控体系；

2.详细构建过程

其中我的promethues的配置如下：

#全局配置信息：
global:
  scrape_interval:     15s #默认抓取间隔1m, 15秒向目标抓取一次数据
  evaluation_interval: 15s 

#个性化的抓取配置信息
scrape_configs:
  - job_name: prometheus
    static_configs:
      - targets: ['192.168.32.70:9090']
        labels:
          instance:  192.168.32.70
 
#监控物理主机：cpu、memory、disk等 
  - job_name: node_exporter
    static_configs:
      - targets:
        - '192.168.32.62:9100'
        - '192.168.32.63:9100'
        - '192.168.32.64:9100'
        - '192.168.32.65:9100'
        - '192.168.32.66:9100'
        - '192.168.32.67:9100'
        - '192.168.32.68:9100'
        - '192.168.32.69:9100'
        - '192.168.32.70:9100'
        - '192.168.32.71:9100'
        - '192.168.32.72:9100'
        - '192.168.32.73:9100'
        - '192.168.32.101:9100'
        - '192.168.32.106:9100'
        - '192.168.32.111:9100'
        - '192.168.32.116:9100'
        - '192.168.32.121:9100'
        - '192.168.32.126:9100'
        - '192.168.32.141:9100'
        - '192.168.32.146:9100'


#  - job_name: 'redis'
#    static_configs:
#     - targets: ['192.168.32.70:9121']
#       labels:
#         instance: redis-poc

  - job_name: 'mongo'
    static_configs:
     - targets: ['192.168.32.70:9104']
       labels:
         instance: mongo3


  - job_name: 'postgres'
    static_configs:
     - targets: ['192.168.32.70:9187']
       labels:
         instance: 'node70'
         platform: 'dcos-node70'

  - job_name: 'kafka'
    static_configs:
     - targets: ['192.168.32.70:9308']
       labels:
         instance: kafka-67


  - job_name: 'redis_exporter_poc'
    static_configs:
      - targets:
        - redis://192.168.32.141:16055
        - redis://192.168.32.68:16052
        - redis://192.168.32.68:16053
    metrics_path: /scrape
    relabel_configs:
      - source_labels: [__address__]
        target_label: __param_target
      - source_labels: [__param_target]
        target_label: instance
      - target_label: __address__
        replacement: 192.168.32.70:9121


  - job_name: 'elasticsearch'
    scrape_interval: 15s
    metrics_path: "/_prometheus/metrics"
    static_configs:
    - targets: ['192.168.32.62:9200']

---

3.构建之后的效果

grafana指标展示界面

2.在线下载已知的控制面板

监控虚拟机物理节点资源

kafka相关监控体系

其他的类似，还有redis、postgresql、mysql、MongoDB等相关的监控体系；

甚至还有微服务相关的监控体系

时间关系，大概展示一些图片，其实dashboard面板还是挺美观的

---