Prometheus 监控之 zookeeper

末蓝、 2022-04-15 06:38 574阅读 0赞

#### zookeeper 监控 ####

git项目地址：[https://github.com/jiankunking/zookeeper\_exporter][https_github.com_jiankunking_zookeeper_exporter]  
exporter下载地址：[https://github.com/carlpett/zookeeper\_exporter/releases/download/v1.0.2/zookeeper\_exporter][https_github.com_carlpett_zookeeper_exporter_releases_download_v1.0.2_zookeeper_exporter]

[sss@keeper01 ~]$ /usr/local/bin/zookeeper_exporter --help
    Usage of /usr/local/bin/zookeeper_exporter:
      -bind-addr string
            bind address for the metrics server (default ":9141")
      -log-level string
            log level (default "info")
      -metrics-path string
            path to metrics endpoint (default "/metrics")
      -reset-on-scrape
            should a reset command be sent to zookeeper on each scrape (default true)
      -version
            show version and exit
      -zookeeper string
            host:port for zookeeper socket (default "localhost:2181")
    [sss@keeper01 ~]$ /usr/local/bin/zookeeper_exporter
    [sss@keeper01 ~]$ curl localhost:9141/metrics 
    # HELP go_gc_duration_seconds A summary of the GC invocation durations.
    # TYPE go_gc_duration_seconds summary
    go_gc_duration_seconds{ quantile="0"} 3.1513e-05
    go_gc_duration_seconds{ quantile="0.25"} 4.2555e-05
    go_gc_duration_seconds{ quantile="0.5"} 4.9278e-05
    go_gc_duration_seconds{ quantile="0.75"} 8.2042e-05
    go_gc_duration_seconds{ quantile="1"} 0.000212331
    go_gc_duration_seconds_sum 0.002286821
    go_gc_duration_seconds_count 31
    # HELP go_goroutines Number of goroutines that currently exist.
    # TYPE go_goroutines gauge
    go_goroutines 14
    ……

ZooKeeper 提供了四字命令(The Four Letter Words)，用来获取 ZooKeeper 服务的当前状态及相关信息。

有哪些命令可以使用？

ZooKeeper四字命令	功能描述
    conf	打印配置
    cons	列出所有连接到这台服务器的客户端全部连接/会话详细信息。包括"接受/发送"的包数量、会话id、操作延迟、最后的操作执行等等信息。
    crst	重置所有连接的连接和会话统计信息。
    dump	列出那些比较重要的会话和临时节点。这个命令只能在leader节点上有用。
    envi	打印出服务环境的详细信息。
    reqs	列出未经处理的请求
    ruok	即"Are you ok"，测试服务是否处于正确状态。如果确实如此，那么服务返回"imok"，否则不做任何相应。
    stat	输出关于性能和连接的客户端的列表。
    srst	重置服务器的统计。
    srvr	列出连接服务器的详细信息
    wchs	列出服务器watch的详细信息。
    wchc	通过session列出服务器watch的详细信息，它的输出是一个与watch相关的会话的列表。
    wchp	通过路径列出服务器watch的详细信息。它输出一个与session相关的路径。
    mntr	输出可用于检测集群健康状态的变量列表

可以在客户端可以通过 telnet 或 nc 向 ZooKeeper 提交相应的命令。举个最常用的栗子：

echo mntr | nc ip 2181

###### 指标名 解释 ######

<table> 
 <thead> 
  <tr> 
   <th>指标名</th> 
   <th>解释</th> 
  </tr> 
 </thead> 
 <tbody> 
  <tr> 
   <td>zk_version</td> 
   <td>版本</td> 
  </tr> 
  <tr> 
   <td>zk_avg_latency</td> 
   <td>平均 响应延迟</td> 
  </tr> 
  <tr> 
   <td>zk_max_latency</td> 
   <td>最大 响应延迟</td> 
  </tr> 
  <tr> 
   <td>zk_min_latency</td> 
   <td>最小 响应延迟</td> 
  </tr> 
  <tr> 
   <td>zk_packets_received</td> 
   <td>收包数</td> 
  </tr> 
  <tr> 
   <td>zk_packets_sent</td> 
   <td>发包数</td> 
  </tr> 
  <tr> 
   <td>zk_num_alive_connections</td> 
   <td>活跃连接数</td> 
  </tr> 
  <tr> 
   <td>zk_outstanding_requests</td> 
   <td>堆积请求数</td> 
  </tr> 
  <tr> 
   <td>zk_server_state</td> 
   <td>主从状态</td> 
  </tr> 
  <tr> 
   <td>zk_znode_count</td> 
   <td>znode 数</td> 
  </tr> 
  <tr> 
   <td>zk_watch_count</td> 
   <td>watch 数</td> 
  </tr> 
  <tr> 
   <td>zk_ephemerals_count</td> 
   <td>临时节点数</td> 
  </tr> 
  <tr> 
   <td>zk_approximate_data_size</td> 
   <td>近似数据总和大小</td> 
  </tr> 
  <tr> 
   <td>zk_open_file_descriptor_count</td> 
   <td>打开 文件描述符 数</td> 
  </tr> 
  <tr> 
   <td>zk_max_file_descriptor_count</td> 
   <td>最大 文件描述符 数</td> 
  </tr> 
  <tr> 
   <td>leader才有的指标</td> 
   <td></td> 
  </tr> 
  <tr> 
   <td>zk_followers</td> 
   <td>Follower 数</td> 
  </tr> 
  <tr> 
   <td>zk_synced_followers</td> 
   <td>已同步的 Follower 数</td> 
  </tr> 
  <tr> 
   <td>zk_pending_syncs</td> 
   <td>阻塞中的 sync 操作</td> 
  </tr> 
 </tbody> 
</table>

##### 需要指定阈值的指标 #####

zk\_outstanding\_requests 堆积请求数  
zk\_pending\_syncs 阻塞中的 sync 操作  
zk\_avg\_latency 平均 响应延迟  
zk\_open\_file\_descriptor\_count 打开 文件描述符 数  
zk\_max\_file\_descriptor\_count 最大 文件描述符 数  
zk\_up 1  
zk\_server\_state 主从状态  
zk\_num\_alive\_connections 活跃连接数

##### rule文件（仅供参考）： #####

groups:
    - name: zookeeperStatsAlert 
      rules:
      - alert: 堆积请求数过大
        expr: avg(zk_outstanding_requests) by (instance) > 10 
        for: 1m
        labels:
          severity: critical
        annotations:
          summary: "Instance { { $labels.instance }} "
          description: "积请求数过大"
      - alert: 阻塞中的 sync 过多
        expr: avg(zk_pending_syncs) by (instance) > 10 
        for: 1m
        labels:
          severity: critical
        annotations:
          summary: "Instance { { $labels.instance }} " 
          description: "塞中的 sync 过多"
      - alert: 平均响应延迟过高
        expr: avg(zk_avg_latency) by (instance) > 10
        for: 1m
        labels:
          severity: critical
        annotations:
          summary: "Instance { { $labels.instance }} "
          description: '平均响应延迟过高'
      - alert: 打开文件描述符数大于系统设定的大小
        expr: zk_open_file_descriptor_count > zk_max_file_descriptor_count * 0.85
        for: 1m
        labels:
          severity: critical
        annotations:
          summary: "Instance { { $labels.instance }} "
          description: '打开文件描述符数大于系统设定的大小'
      - alert: zookeeper服务器宕机
        expr: zk_up == 0
        for: 1m
        labels:
          severity: critical
        annotations:
          summary: "Instance { { $labels.instance }} "
          description: 'zookeeper服务器宕机'
      - alert: zk主节点丢失
        expr: absent(zk_server_state{ state="leader"})  != 1 
        for: 1m
        labels:
          severity: critical
        annotations:
          summary: "Instance { { $labels.instance }} "
          description: 'zk主节点丢失'

##### Grafana画图： #####

在Grafana分享了zk监控图：搜索Zookeeper Exporer Overview 或者 拷贝pid 9236  
![在这里插入图片描述][watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3FxXzI1OTM0NDAx_size_16_color_FFFFFF_t_70]

#### kafka 监控 ####

git项目地址：[https://github.com/danielqsj/kafka\_exporter][https_github.com_danielqsj_kafka_exporter]  
下载地址： [https://github.com/danielqsj/kafka\_exporter/releases/download/v1.2.0/kafka\_exporter-1.2.0.linux-amd64.tar.gz][https_github.com_danielqsj_kafka_exporter_releases_download_v1.2.0_kafka_exporter-1.2.0.linux-amd64.tar.gz]

启动

kafka_exporter --kafka.server=kafka:9092 [--kafka.server=another-server ...]

[https_github.com_jiankunking_zookeeper_exporter]: https://github.com/jiankunking/zookeeper_exporter
[https_github.com_carlpett_zookeeper_exporter_releases_download_v1.0.2_zookeeper_exporter]: https://github.com/carlpett/zookeeper_exporter/releases/download/v1.0.2/zookeeper_exporter
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3FxXzI1OTM0NDAx_size_16_color_FFFFFF_t_70]: /images/20220415/331db7f976bc4e67ad8c17e03c5197aa.png
[https_github.com_danielqsj_kafka_exporter]: https://github.com/danielqsj/kafka_exporter
[https_github.com_danielqsj_kafka_exporter_releases_download_v1.2.0_kafka_exporter-1.2.0.linux-amd64.tar.gz]: https://github.com/danielqsj/kafka_exporter/releases/download/v1.2.0/kafka_exporter-1.2.0.linux-amd64.tar.gz

Prometheus 监控之 zookeeper

发表评论取消回复

还没有评论，来说两句吧...

相关阅读