博客

prometheus使用 (二十) 监控Kubernetes集群

  • Post author:
  • Post category:其他



之前我们都是直接去抓取部署在宿主机上服务暴露出来的指标进行监控


K8是什么就不说了，本章主要是说一下在宿主机和容器中prometheus该如何监控K8S集群

一、宿主机prometheus监听K8


其实比较常见的还是在容器中跑prometheus，但在业务中也碰到过在宿主机部署的情况

1、监听node资源

vi /etc/prometheus/prometheus.yml 

global:
  scrape_interval:     1m
  evaluation_interval: 30s

scrape_configs:
- job_name: test       
  scrape_interval:    30s
  scheme: https    
  tls_config:      
    insecure_skip_verify: true
    ca_file: /etc/kubernetes/pki/ca.crt
    cert_file:  /etc/kubernetes/pki/apiserver-kubelet-client.crt
    key_file:  /etc/kubernetes/pki/apiserver-kubelet-client.key
  kubernetes_sd_configs:        
  - role: node                
    api_server: https://10.0.16.16:6443    
    tls_config:                     
      ca_file: /etc/kubernetes/pki/ca.crt
      cert_file:  /etc/kubernetes/pki/apiserver-kubelet-client.crt
      key_file:  /etc/kubernetes/pki/apiserver-kubelet-client.key


配置说明
 

scrape_configs:
- job_name: test       
  scrape_interval:    30s
  scheme: https             #这个https是表示我们要请求/metrics 
                            #这个链接是https模式的



  tls_config:    #认证部分，意思是我们采集https://xxx/metrics这个接口时的提供认证
                 #因为是监听K8S，指标一般也是走K8s的接口的，所以证书直接用集群的就行
    insecure_skip_verify: true              
    ca_file: /etc/kubernetes/pki/ca.crt  
    cert_file:  /etc/kubernetes/pki/apiserver-kubelet-client.crt
    key_file:  /etc/kubernetes/pki/apiserver-kubelet-client.key
                            
                           

  kubernetes_sd_configs:    #K8的服务发现
  - role: node              #要获取的资源是什么
    api_server: https://10.0.16.16:6443    #K8S集群master地址
                                           #我们要从这个地址的接口请求指标

    tls_config:                            #请求K8S集群的认证文件
      ca_file: /etc/kubernetes/pki/ca.crt
      cert_file:  /etc/kubernetes/pki/apiserver-kubelet-client.crt
      key_file:  /etc/kubernetes/pki/apiserver-kubelet-client.key

启动服务 

systemctl restart prometheus


这里说明一下，我这次用的是云主机。只有1台所以即是master也是node


图中找到的是node主机上kubelet10250端口，如果不明白可以用下面的命令去手动请求下
 

curl  https://10.0.16.16:10250/metrics \
--cacert /etc/kubernetes/pki/ca.crt  \
--cert /etc/kubernetes/pki/apiserver-kubelet-client.crt \
--key /etc/kubernetes/pki/apiserver-kubelet-client.key \
-k


上面就是”node角色” 所监听的所有指标，当然K8中也不仅仅只有node一种资源


还有ep、ingres、svc等等，我们都要通过 kubernetes_sd_configs 去做服务发现



kubernetes_sd_config · 笔记 · 看云

可以参考这位大大的说明

2、监听集群master资源


我们平时使用K8如果想要获取精准的master地址，通常会使用如下命令
 

[root@VM-16-16-centos setup]# kubectl -n default get endpoints kubernetes
NAME         ENDPOINTS         AGE
kubernetes   10.0.16.16:6443   19d


在kubernetes_sd_configs服务发现中有几个类似的内置指标
 

#别说找不到，上面刚刚分享的链接里都有
__meta_kubernetes_namespace           #服务对象的命名空间。
__meta_kubernetes_service_name        #服务对象的名称。
__meta_kubernetes_endpoint_port_name  #端点端口的名称  (是否是https打头)


编写配置

vi /etc/prometheus/prometheus.yml 

- job_name: apiserver
  scrape_interval:    30s
  scheme: https
  tls_config:
    insecure_skip_verify: true
    ca_file: /etc/kubernetes/pki/ca.crt
    cert_file:  /etc/kubernetes/pki/apiserver-kubelet-client.crt
    key_file:  /etc/kubernetes/pki/apiserver-kubelet-client.key
  kubernetes_sd_configs:
  - role: endpoints     
    api_server: https://10.0.16.16:6443
    tls_config:
      ca_file: /etc/kubernetes/pki/ca.crt
      cert_file:  /etc/kubernetes/pki/apiserver-kubelet-client.crt
      key_file:  /etc/kubernetes/pki/apiserver-kubelet-client.key
  relabel_configs:
  - source_labels: 
      - __meta_kubernetes_namespace
      - __meta_kubernetes_service_name
      - __meta_kubernetes_endpoint_port_name
    regex: default;kubernetes;https
    action: keep


配置说明
 

  kubernetes_sd_configs:
  - role: endpoints       #我们上面kubectl get的资源是endpoint
                          #说明这个指标来自于endpoints这个资源对象

    api_server: https://10.0.16.16:6443     #同上node资源
    tls_config:
      ca_file: /etc/kubernetes/pki/ca.crt
      cert_file:  /etc/kubernetes/pki/apiserver-kubelet-client.crt
      key_file:  /etc/kubernetes/pki/apiserver-kubelet-client.key


  relabel_configs:    #标签重写，上一章的知识点
  - source_labels:       #要重新的标签是啥
      - __meta_kubernetes_namespace            #命名空间地址
      - __meta_kubernetes_service_name         #svc的名称，一般和ep是一样的
      - __meta_kubernetes_endpoint_port_name   #指标是否是https模式
   
    regex: default;kubernetes;https  #按照顺序定义上面3个标签的值
    action: keep                     #将所有指标中，不匹配这3个元标签值的指标都丢弃掉

重载配置 

curl -XPOST http://10.0.16.16:9090/-/reload


在服务发现中我们可以看到他已经将不满足上面3个标签的指标都丢弃了


容器中prometheus的配置和宿主机配置 基本相同，只需要注意一下证书问题


我们下面则以容器的形式去部署prometheus，来监控整个K8集群的资源

二、容器prometheus监控集群

1、部署prometheus容器


防止冲突，这里先将宿主机的prometheus服务停掉
 

systemctl stop prometheus


添加prometheusyaml(直接复制即可）
 

cat > prometheus-dev.yaml  <<EOF
---   
apiVersion: rbac.authorization.k8s.io/v1
kind: ClusterRole
metadata:
  name: prometheus
rules:
- apiGroups: [""]
  resources:
  - nodes
  - nodes/proxy
  - services
  - endpoints
  - pods
  verbs: ["get", "list", "watch"]
- apiGroups:
  - extensions
  resources:
  - ingresses
  verbs: ["get", "list", "watch"]
- nonResourceURLs: ["/metrics"]
  verbs: ["get"]
---
apiVersion: v1
kind: ServiceAccount
metadata:
  name: prometheus
  namespace: kube-system
---
apiVersion: rbac.authorization.k8s.io/v1
kind: ClusterRoleBinding
metadata:
  name: prometheus
roleRef:
  apiGroup: rbac.authorization.k8s.io
  kind: ClusterRole
  name: prometheus
subjects:
- kind: ServiceAccount
  name: prometheus
  namespace: kube-system
---
kind: Service
apiVersion: v1
metadata:
  labels:
    app: prometheus
  name: prometheus
  namespace: kube-system
spec:
  ports:
  - port: 9090
    targetPort: 9090
  selector:
    app: prometheus
---
apiVersion: apps/v1
kind: Deployment
metadata:
  labels:
    name: prometheus-deployment
  name: prometheus
  namespace: kube-system
spec:
  replicas: 1
  selector:
    matchLabels:
      app: prometheus
  template:
    metadata:
      labels:
        app: prometheus
    spec:
      containers:
      - image: prom/prometheus
        name: prometheus
        command:
        - "/bin/prometheus"
        args:
        - "--web.enable-lifecycle"   #可以curl重载配置了
        - "--config.file=/etc/prometheus/prometheus.yml"
        - "--storage.tsdb.path=/prometheus"
        - "--storage.tsdb.retention=24h"
        ports:
        - containerPort: 9090
          protocol: TCP
          hostPort: 9090
        volumeMounts:
        - mountPath: "/prometheus"
          name: data
        - mountPath: "/etc/prometheus"
          name: config-volume
        resources:
          requests:
            cpu: 100m
            memory: 100Mi
          limits:
            cpu: 500m
            memory: 2500Mi
      serviceAccountName: prometheus    
      volumes:
      - name: data
        emptyDir: {}
      - name: config-volume
        configMap:
          name: prometheus-config
EOF


添加configmap配置
 

cat > cm.yaml  <<EOF
apiVersion: v1
kind: ConfigMap
metadata:
  name: prometheus-config
  namespace: kube-system
data:
  prometheus.yml: |
    global:
      scrape_interval:     15s
      evaluation_interval: 15s
EOF


部署容器
 

kubectl apply -f cm.yaml 
kubectl apply -f prometheus-dev.yaml

2、监控prometheus本身

vi cm.yml 

    scrape_configs:        
    - job_name: 'prometheus'
      scrape_interval: 5s
      static_configs:
      - targets: ['localhost:9090']   #这个没什么可说的，添加监控自身，也可以不要

重载 

kubectl apply -f cm.yaml
curl -XPOST http://10.0.16.16:9090/-/reload  #有时候会很慢，可以多跑几下

3、监控apiserver端点


这个我们上面宿主机已经写过了，直接复制过来修改一下证书

vi cm.yaml 

    - job_name: apiserver
      scrape_interval: 30s
      scheme: https
      kubernetes_sd_configs:
      - role: endpoints
      tls_config:     
        ca_file: /var/run/secrets/kubernetes.io/serviceaccount/ca.crt
        insecure_skip_verify: true
      bearer_token_file: /var/run/secrets/kubernetes.io/serviceaccount/token
      relabel_configs:
      - source_labels: [__meta_kubernetes_namespace,__meta_kubernetes_service_name,__meta_kubernetes_endpoint_port_name]
        action: keep
        regex: default;kubernetes;https


配置说明
 

    - job_name: apiserver
      scrape_interval: 30s
      scheme: https         #因为在集群内部，这里就不用写请求指标的证书了
      kubernetes_sd_configs:
      - role: endpoints
      tls_config:          #K8是有自动注入的证书的，直接复制下面的使用即可，不要纠结，通用~
        ca_file: /var/run/secrets/kubernetes.io/serviceaccount/ca.crt
        insecure_skip_verify: true
      bearer_token_file: /var/run/secrets/kubernetes.io/serviceaccount/token
      relabel_configs:
      - source_labels: [__meta_kubernetes_namespace,__meta_kubernetes_service_name,__meta_kubernetes_endpoint_port_name]
        action: keep
        regex: default;kubernetes;https

重载配置 

kubectl apply -f cm.yaml
curl -XPOST http://10.0.16.16:9090/-/reload  

#如果实在刷不出来，就把prometheus的pod干掉，效果一样

4、监听kubelet组件


简单来说，kubelet有个端口10250、10255 是用来暴露节点metrics指标的


但是后面不怎么用了，改用内部接口了  接口形式如下
 

/api/v1/nodes/主机名/proxy/metrics

vi cm.yaml 

    - job_name: kubernetes-nodes-kubelet-containers
      kubernetes_sd_configs:
      - role: node    #上面我们得知kubelet的指标是在api/v1/nodes下的，所以用node角色
      scheme: https
      tls_config:
        ca_file: /var/run/secrets/kubernetes.io/serviceaccount/ca.crt
        insecure_skip_verify: true
      bearer_token_file: /var/run/secrets/kubernetes.io/serviceaccount/token

部署 

kubectl apply -f cm.yaml
curl -XPOST http://10.0.16.16:9090/-/reload


思考


奇奇怪怪，为什么宿主机上这样可以用，K8里面就没权限了呢？


有没有可能是因为prometheus是从容器集群内部去访问集群外部的kubelet接口

然而，外边kubelet不认K8注入的证书呢  如果我们再去外部拿证书又很麻烦


ψ(｀∇´)ψ


解决方法


偶然间看到 kubelet的指标是可以通过apiserver的接口去获取的。如下:

/api/v1/nodes/主机名/proxy/metrics
 

#那么我们就要想办法去将抓取指标的操作指定到master的这个接口上
#解决思路

1、首先我们先通过修改"__address__"内置标签默认的值10.0.16.16:10250 ,指定到我们的master地址
2、想办法找到 node的主机名是什么，正好我们上面kubernetes_sd_configs中有这么一条
         __meta_kubernetes_node_name  可以获取节点的主机名
3、将prometheus默认抓取的路径/metrics 修改为/api/v1/nodes/主机名/proxy/metrics 的接口路径


添加重写配置
 

      relabel_configs:    #重写标签部分
      - target_label: __address__                   
        replacement: kubernetes.default.svc:443  #也可以直接指master地址+端口

      - source_labels: [__meta_kubernetes_node_name]  
        regex: (.+)                    #将匹配到的主机名通过${1} 调用
        target_label: __metrics_path__                 #替换获取metrics的路径
        replacement: /api/v1/nodes/${1}/proxy/metrics  #准备替换的metrics路径


重载配置
 

kubectl apply -f cm.yaml
curl -XPOST http://10.0.16.16:9090/-/reload


图中可以看到，我们通过集群内部接口api获取到了主机的kubelet指标



https://kubernetes.default.svc/api/v1/nodes/vm-16-16-centos/proxy/metrics


当然我们也可以在指定kubernetes.default.svc:443时修改为master地址也是可行的


5、监听cadvisor组件


cadvisor已经被集成在kebelet组件中了，采集和和kubelet类似，接口为


/api/v1/nodes/主机名/proxy/metrics/cadvisor

vi cm.yaml 

    - job_name: 'kubernetes-cadvisor'
      kubernetes_sd_configs:
      - role: node
      scheme: https
      tls_config:
        ca_file: /var/run/secrets/kubernetes.io/serviceaccount/ca.crt
      bearer_token_file: /var/run/secrets/kubernetes.io/serviceaccount/token
      relabel_configs:
      - action: labelmap
        regex: __meta_kubernetes_node_label_(.+)   #把该node所有标签都列出来
      - target_label: __address__
        replacement: kubernetes.default.svc:443
      - source_labels: [__meta_kubernetes_node_name]
        regex: (.+)
        target_label: __metrics_path__
        replacement: /api/v1/nodes/${1}/proxy/metrics/cadvisor  #区别在于接口不同

重载配置 

kubectl apply -f cm.yaml
curl -XPOST http://10.0.16.16:9090/-/reload

6、监控 node-exporter

cat > node-exporter.yaml  << EOF
apiVersion: apps/v1
kind: DaemonSet
metadata:
  name: node-exporter
  namespace: kube-system
  labels:
    k8s-app: node-exporter
spec:
  selector:
    matchLabels:
      k8s-app: node-exporter
  template:
    metadata:
      labels:
        k8s-app: node-exporter
    spec:
      containers:
      - image: prom/node-exporter:v1.0.1
        name: node-exporter
        ports:
        - containerPort: 9100
          protocol: TCP
          name: http
          hostPort: 9100
        volumeMounts:
        - mountPath: /root-disk
          name: root-disk
          readOnly: true
        - mountPath: /docker-disk
          name: docker-disk
          readOnly: true
      volumes:
      - name: root-disk
        hostPath:
          path: /
          type: ""
      - name: docker-disk
        hostPath:
          path: /var/lib/docker 
          type: ""
---
apiVersion: v1
kind: Service
metadata:
  labels:
    k8s-app: node-exporter
  name: node-exporter
  namespace: kube-system
spec:
  ports:
  - name: http
    port: 9100
    protocol: TCP
  selector:
    k8s-app: node-exporter
EOF

部署 

kubectl apply -f node-exporter.yaml


添加监控配置

vi cm.yaml 

    - job_name: 'kubernetes-nodes'
      scrape_interval: 30s
      scrape_timeout: 10s
      kubernetes_sd_configs:
      - role: node
      scheme: http
      tls_config:
        ca_file: /var/run/secrets/kubernetes.io/serviceaccount/ca.crt
      bearer_token_file: /var/run/secrets/kubernetes.io/serviceaccount/token
      relabel_configs:
      - source_labels: [__address__]
        regex: '(.*):10250'    #这里匹配每个节点的ip
        replacement: '${1}:9100'  #然后将ip拿过来，加上9100端口(node-exporter暴露到主机上了)
        target_label: __address__     #替换掉拿数据的地址，就直接指定到node-exporter上了
        action: replace

重载配置 

kubectl apply -f cm.yaml
curl -XPOST http://10.0.16.16:9090/-/reload


我是 单节点，你们的会多一些是( •̀ ω •́ )✧

完整configmap配置如下

apiVersion: v1
kind: ConfigMap
metadata:
  name: prometheus-config
  namespace: kube-system
data:
  prometheus.yml: |
    global:
      scrape_interval:     15s
      evaluation_interval: 15s
    scrape_configs:        
    - job_name: 'prometheus'
      scrape_interval: 5s
      static_configs:
      - targets: ['localhost:9090']   #这个没什么可说的，添加监控自身，也可以不要
    - job_name: apiserver
      scrape_interval: 30s
      scheme: https
      kubernetes_sd_configs:
      - role: endpoints
      tls_config:     
        ca_file: /var/run/secrets/kubernetes.io/serviceaccount/ca.crt
        insecure_skip_verify: true
      bearer_token_file: /var/run/secrets/kubernetes.io/serviceaccount/token
      relabel_configs:
      - source_labels: [__meta_kubernetes_namespace,__meta_kubernetes_service_name,__meta_kubernetes_endpoint_port_name]
        action: keep
        regex: default;kubernetes;https
    - job_name: kubernetes-nodes-kubelet-containers
      kubernetes_sd_configs:
      - role: node    #上面我们得知kubelet的指标是在api/v1/nodes下的，所以用node角色
      scheme: https
      tls_config:
        ca_file: /var/run/secrets/kubernetes.io/serviceaccount/ca.crt
        insecure_skip_verify: true
      bearer_token_file: /var/run/secrets/kubernetes.io/serviceaccount/token
      relabel_configs:    #重写标签部分


      - target_label: __address__                   
        replacement: kubernetes.default.svc:443

      - source_labels: [__meta_kubernetes_node_name]  
        regex: (.+)
        target_label: __metrics_path__  
        replacement: /api/v1/nodes/${1}/proxy/metrics 
    - job_name: 'kubernetes-cadvisor'
      kubernetes_sd_configs:
      - role: node
      scheme: https
      tls_config:
        ca_file: /var/run/secrets/kubernetes.io/serviceaccount/ca.crt
      bearer_token_file: /var/run/secrets/kubernetes.io/serviceaccount/token
      relabel_configs:
      - action: labelmap
        regex: __meta_kubernetes_node_label_(.+)  #这个操作和kubelet基本一致
      - target_label: __address__
        replacement: kubernetes.default.svc:443
      - source_labels: [__meta_kubernetes_node_name]
        regex: (.+)
        target_label: __metrics_path__
        replacement: /api/v1/nodes/${1}/proxy/metrics/cadvisor  #区别在于接口不同
    - job_name: 'kubernetes-nodes'
      scrape_interval: 30s
      scrape_timeout: 10s
      kubernetes_sd_configs:
      - role: node
      scheme: http
      tls_config:
        ca_file: /var/run/secrets/kubernetes.io/serviceaccount/ca.crt
      bearer_token_file: /var/run/secrets/kubernetes.io/serviceaccount/token
      relabel_configs:
      - source_labels: [__address__]
        regex: '(.*):10250'    #这里匹配每个节点的ip
        replacement: '${1}:9100'  #然后将ip拿过来，加上9100端口(node-exporter暴露到主机上了)
        target_label: __address__     #替换掉拿数据的地址，就直接指定到node-exporter上了
        action: replace


版权声明：本文为qq_42883074原创文章，遵循 CC 4.0 BY-SA 版权协议，转载请附上原文出处链接和本声明。