部署k8s集群监控Heapster

git clone https://github.com/kubernetes/heapster.git
kubectl apply -f heapster/deploy/kube-config/influxdb/
kubectl apply -f heapster/deploy/kube-config/rbac/heapster-rbac.yaml
#通过上面yaml部署之后，会报无法拉去镜像k8s.gcr.io，找到相关镜像并tag（docker search到的）

#修改monitoring-grafana这个svc的Type为NodePort或者新建一个ingress以暴露服务供外界访问
kubectl edit svc monitoring-grafana



遇到的问题

1、heapster pod中报如下错误（10255端口不通，k8s默认使用10250作为kubelet端口）
E0730 11:24:05.015741       1 manager.go:101] Error in scraping containers from kubelet:172.31.2.131:10255: failed to get all container stats from Kubelet URL "http://172.31.2.131:10255/stats/container/": Post http://172.31.2.131:10255/stats/container/: dial tcp 172.31.2.131:10255: getsockopt: connection refused
解决：
kubectl edit deployment heapster

 修改

 - --source=kubernetes:https://kubernetes.default

 为

 - --source=kubernetes:kubernetes:https://kubernetes.default?useServiceAccount=true&kubeletHttps=true&kubeletPort=10250&insecure=true

 

  2. heapster pod中报如下错误（rbac权限不够，但是看了heapster这个clusterrolebinding，发现是有足够的权限get list watch pod的）

  E0823 02:26:05.018478 1 kubelet.go:239] error while getting containers from Kubelet 172.16.0.12:10250: failed to get all container stats from Kubelet URL "https://172.16
  .0.12:10250/stats/container/": request failed - "403 Forbidden", response: "Forbidden (user=system:serviceaccount:kube-system:heapster, verb=create, resource=nodes, subresourc
  e=stats)"

  临时解决：删除crb  kubectl delete clusterrolebinding heapster

          新建一个绑定更高clusterrole的crb并重启pod

          

apiVersion: rbac.authorization.k8s.io/v1
kind: ClusterRoleBinding
metadata:
  name: heapster
roleRef:
  apiGroup: rbac.authorization.k8s.io
  kind: ClusterRole
  name: cluster-admin #cluster-admin的权限
subjects:
- kind: ServiceAccount
  name: heapster
  namespace: kube-system

 

 

验证

通过使用nodeIp：NodePort访问服务，可以看到默认的Cluster和Pod统计图

使用kubectl top node如下

[root@node1 kubelet.service.d]# kc top node
NAME        CPU(cores)   CPU%      MEMORY(bytes)   MEMORY%   
node1.com   175m         8%        2285Mi          61%       
node2.com   97m          4%        2094Mi          56%      

 

[root@node1 kubelet.service.d]# kc top pod
NAME                                        CPU(cores)   MEMORY(bytes)   
altered-horse-mysql-5fd7bcc98c-9sc7n        2m           241Mi           
austere-llama-mariadb-7f578657c6-9rd84      3m           143Mi           
austere-llama-wordpress-75cbff959-tz469     3m           223Mi           
coredns-78fcdf6894-5bwc7                    1m           11Mi            
coredns-78fcdf6894-x79ww                    1m           12Mi            
default-http-backend-784cf5f7b5-p9mcd       0m           2Mi             
etcd-node1.com                              16m          186Mi           
heapster-798c4b6988-xmnzp                   2m           51Mi            
kube-apiserver-node1.com                    25m          306Mi           
kube-controller-manager-node1.com           28m          64Mi            
kube-flannel-ds-8zcwd                       1m           16Mi            
kube-flannel-ds-gldqx                       1m           18Mi            
kube-proxy-szb9p                            3m           15Mi            
kube-proxy-tq7d7                            3m           15Mi            
kube-scheduler-node1.com                    8m           19Mi            
monitoring-grafana-77fbf95454-4k752         0m           15Mi            
monitoring-influxdb-56b597d5c5-55kpf        2m           65Mi            
nginx-ingress-controller-84b8b6df66-46848   3m           138Mi           
tiller-deploy-7b5cb97898-dnmc2              0m           15Mi