Kubernetes监控体系（三）之Heapster介绍

心已赠人 2022-06-08 00:23 204阅读 0赞

> 本文个人博客地址：[http://www.huweihuang.com/article/kubernetes/monitoring/heapster-introduction/][http_www.huweihuang.com_article_kubernetes_monitoring_heapster-introduction]

## 1. heapster简介 ##

Heapster是容器集群监控和性能分析工具，支持Kubernetes和CoreOS。  
Kubernetes有个监控agent—`cAdvisor`。在每个kubernetes Node上都会运行cAdvisor，它会收集本机以及容器的监控数据(cpu,memory,filesystem,network,uptime)。

## 2. heapster部署与配置 ##

### 2.1. 注意事项 ###

需同步部署机器和被采集机器的时间：ntpdate time.windows.com

加入定时任务，定期同步时间

crontab –e

30 5 \* \* \* /usr/sbin/ntpdate time.windows.com //每天早晨5点半执行

### 2.2. 容器式部署 ###

以下为示例，请拉取官方镜像，根据实际情况修改参数。

#拉取镜像
    docker pull heapster:latest
    #运行容器
    docker run -d -p 8082:8082 --net=host heapster:latest --source=kubernetes:http://<k8s-server-ip>:8080?inClusterConfig=false\&useServiceAccount=false --sink=influxdb:http://<influxdb-ip>:8086?db=<k8s_env_zone>

### 2.3. 配置说明 ###

可以参考[官方文档][Link 1]

#### 2.3.1. **–source** ####

–source: 指定数据获取源。这里指定kube-apiserver即可。  
后缀参数：  
inClusterConfig:  
kubeletPort: 指定kubelet的使用端口，默认10255  
kubeletHttps: 是否使用https去连接kubelets(默认：false)  
apiVersion: 指定K8S的apiversion  
insecure: 是否使用安全证书(默认：false)  
auth: 安全认证  
useServiceAccount: 是否使用K8S的安全令牌

#### 2.3.2. **–sink** ####

–sink: 指定后端数据存储。这里指定influxdb数据库。  
后缀参数：  
user: InfluxDB用户  
pw: InfluxDB密码  
db: 数据库名  
secure: 安全连接到InfluxDB(默认：false)  
withfields： 使用InfluxDB fields(默认：false)。

## 3. Metrics ##

<table> 
 <thead> 
  <tr> 
   <th>分类</th> 
   <th>Metric Name</th> 
   <th>Description</th> 
   <th>备注</th> 
  </tr> 
 </thead> 
 <tbody> 
  <tr> 
   <td>cpu</td> 
   <td>cpu/limit</td> 
   <td>CPU hard limit in millicores.</td> 
   <td>CPU上限</td> 
  </tr> 
  <tr> 
   <td></td> 
   <td>cpu/node_capacity</td> 
   <td>Cpu capacity of a node.</td> 
   <td>Node节点的CPU容量</td> 
  </tr> 
  <tr> 
   <td></td> 
   <td>cpu/node_allocatable</td> 
   <td>Cpu allocatable of a node.</td> 
   <td>Node节点可分配的CPU</td> 
  </tr> 
  <tr> 
   <td></td> 
   <td>cpu/node_reservation</td> 
   <td>Share of cpu that is reserved on the node allocatable.</td> 
   <td></td> 
  </tr> 
  <tr> 
   <td></td> 
   <td>cpu/node_utilization</td> 
   <td>CPU utilization as a share of node allocatable.</td> 
   <td></td> 
  </tr> 
  <tr> 
   <td></td> 
   <td>cpu/request</td> 
   <td>CPU request (the guaranteed amount of resources) in millicores.</td> 
   <td></td> 
  </tr> 
  <tr> 
   <td></td> 
   <td>cpu/usage</td> 
   <td>Cumulative CPU usage on all cores.</td> 
   <td>CPU总使用量</td> 
  </tr> 
  <tr> 
   <td></td> 
   <td>cpu/usage_rate</td> 
   <td>CPU usage on all cores in millicores.</td> 
   <td></td> 
  </tr> 
  <tr> 
   <td>filesystem</td> 
   <td>filesystem/usage</td> 
   <td>Total number of bytes consumed on a filesystem.</td> 
   <td>文件系统的使用量</td> 
  </tr> 
  <tr> 
   <td></td> 
   <td>filesystem/limit</td> 
   <td>The total size of filesystem in bytes.</td> 
   <td>文件系统的使用上限</td> 
  </tr> 
  <tr> 
   <td></td> 
   <td>filesystem/available</td> 
   <td>The number of available bytes remaining in a the filesystem</td> 
   <td>可用的文件系统容量</td> 
  </tr> 
  <tr> 
   <td></td> 
   <td>filesystem/inodes</td> 
   <td>The number of available inodes in a the filesystem</td> 
   <td></td> 
  </tr> 
  <tr> 
   <td></td> 
   <td>filesystem/inodes_free</td> 
   <td>The number of free inodes remaining in a the filesystem</td> 
   <td></td> 
  </tr> 
  <tr> 
   <td>memory</td> 
   <td>memory/limit</td> 
   <td>Memory hard limit in bytes.</td> 
   <td>内存上限</td> 
  </tr> 
  <tr> 
   <td></td> 
   <td>memory/major_page_faults</td> 
   <td>Number of major page faults.</td> 
   <td></td> 
  </tr> 
  <tr> 
   <td></td> 
   <td>memory/major_page_faults_rate</td> 
   <td>Number of major page faults per second.</td> 
   <td></td> 
  </tr> 
  <tr> 
   <td></td> 
   <td>memory/node_capacity</td> 
   <td>Memory capacity of a node.</td> 
   <td></td> 
  </tr> 
  <tr> 
   <td></td> 
   <td>memory/node_allocatable</td> 
   <td>Memory allocatable of a node.</td> 
   <td></td> 
  </tr> 
  <tr> 
   <td></td> 
   <td>memory/node_reservation</td> 
   <td>Share of memory that is reserved on the node allocatable.</td> 
   <td></td> 
  </tr> 
  <tr> 
   <td></td> 
   <td>memory/node_utilization</td> 
   <td>Memory utilization as a share of memory allocatable.</td> 
   <td></td> 
  </tr> 
  <tr> 
   <td></td> 
   <td>memory/page_faults</td> 
   <td>Number of page faults.</td> 
   <td></td> 
  </tr> 
  <tr> 
   <td></td> 
   <td>memory/page_faults_rate</td> 
   <td>Number of page faults per second.</td> 
   <td></td> 
  </tr> 
  <tr> 
   <td></td> 
   <td>memory/request</td> 
   <td>Memory request (the guaranteed amount of resources) in bytes.</td> 
   <td></td> 
  </tr> 
  <tr> 
   <td></td> 
   <td>memory/usage</td> 
   <td>Total memory usage.</td> 
   <td></td> 
  </tr> 
  <tr> 
   <td></td> 
   <td>memory/cache</td> 
   <td>Cache memory usage.</td> 
   <td></td> 
  </tr> 
  <tr> 
   <td></td> 
   <td>memory/rss</td> 
   <td>RSS memory usage.</td> 
   <td></td> 
  </tr> 
  <tr> 
   <td></td> 
   <td>memory/working_set</td> 
   <td>Total working set usage. Working set is the memory being used and not easily dropped by the kernel.</td> 
   <td></td> 
  </tr> 
  <tr> 
   <td>network</td> 
   <td>network/rx</td> 
   <td>Cumulative number of bytes received over the network.</td> 
   <td></td> 
  </tr> 
  <tr> 
   <td></td> 
   <td>network/rx_errors</td> 
   <td>Cumulative number of errors while receiving over the network.</td> 
   <td></td> 
  </tr> 
  <tr> 
   <td></td> 
   <td>network/rx_errors_rate</td> 
   <td>Number of errors while receiving over the network per second.</td> 
   <td></td> 
  </tr> 
  <tr> 
   <td></td> 
   <td>network/rx_rate</td> 
   <td>Number of bytes received over the network per second.</td> 
   <td></td> 
  </tr> 
  <tr> 
   <td></td> 
   <td>network/tx</td> 
   <td>Cumulative number of bytes sent over the network</td> 
   <td></td> 
  </tr> 
  <tr> 
   <td></td> 
   <td>network/tx_errors</td> 
   <td>Cumulative number of errors while sending over the network</td> 
   <td></td> 
  </tr> 
  <tr> 
   <td></td> 
   <td>network/tx_errors_rate</td> 
   <td>Number of errors while sending over the network</td> 
   <td></td> 
  </tr> 
  <tr> 
   <td></td> 
   <td>network/tx_rate</td> 
   <td>Number of bytes sent over the network per second.</td> 
   <td></td> 
  </tr> 
  <tr> 
   <td></td> 
   <td>uptime</td> 
   <td>Number of milliseconds since the container was started.</td> 
   <td></td> 
  </tr> 
 </tbody> 
</table>

## 4. Labels ##

<table> 
 <thead> 
  <tr> 
   <th>Label Name</th> 
   <th>Description</th> 
  </tr> 
 </thead> 
 <tbody> 
  <tr> 
   <td>pod_id</td> 
   <td>Unique ID of a Pod</td> 
  </tr> 
  <tr> 
   <td>pod_name</td> 
   <td>User-provided name of a Pod</td> 
  </tr> 
  <tr> 
   <td>pod_namespace</td> 
   <td>The namespace of a Pod</td> 
  </tr> 
  <tr> 
   <td>container_base_image</td> 
   <td>Base image for the container</td> 
  </tr> 
  <tr> 
   <td>container_name</td> 
   <td>User-provided name of the container or full cgroup name for system containers</td> 
  </tr> 
  <tr> 
   <td>host_id</td> 
   <td>Cloud-provider specified or user specified Identifier of a node</td> 
  </tr> 
  <tr> 
   <td>hostname</td> 
   <td>Hostname where the container ran</td> 
  </tr> 
  <tr> 
   <td>labels</td> 
   <td>Comma-separated(Default) list of user-provided labels. Format is ‘key:value’</td> 
  </tr> 
  <tr> 
   <td>namespace_id</td> 
   <td>UID of the namespace of a Pod</td> 
  </tr> 
  <tr> 
   <td>resource_id</td> 
   <td>A unique identifier used to differentiate multiple metrics of the same type. e.x. Fs partitions under filesystem/usage</td> 
  </tr> 
 </tbody> 
</table>

## 5. heapster API ##

见官方文档：[https://github.com/kubernetes/heapster/blob/master/docs/model.md][https_github.com_kubernetes_heapster_blob_master_docs_model.md]

[http_www.huweihuang.com_article_kubernetes_monitoring_heapster-introduction]: http://www.huweihuang.com/article/kubernetes/monitoring/heapster-introduction/
[Link 1]: https://github.com/kubernetes/heapster/tree/master/docs
[https_github.com_kubernetes_heapster_blob_master_docs_model.md]: https://github.com/kubernetes/heapster/blob/master/docs/model.md