K8S学习笔记之安装Metrics-server

艺帆风顺 发布于 2025-04-07 33 次阅读

一、Metrics Server介绍

Metrics Server 是集群级别的资源利用率数据的聚合器。从 Kubelets收集资源指标，并通过 Metrics API 在 Kubernetes apiserver 中公开它们，以供 Horizontal Pod Autoscaler 和Vertical Pod Autoscaler 使用。

Metrics API 也可以通过访问 kubectl top，从而更容易调试自动缩放管道。

Metrics Server基于内存存储，重启后数据将全部丢失，而且它仅能留存最近收集到的指标数据。

二、安装Metrics Server

1、下载配置文件

wget https://github.com/kubernetes-sigs/metrics-server/releases/latest/download/components.yaml

2、修改components.yaml配置文件

修改镜像并新增--kubelet-insecure-tls参数

3、安装

[root@k8s-master ~]# kubectl apply -f components.yaml serviceaccount/metrics-server createdclusterrole.rbac.authorization.k8s.io/system:aggregated-metrics-reader createdclusterrole.rbac.authorization.k8s.io/system:metrics-server createdrolebinding.rbac.authorization.k8s.io/metrics-server-auth-reader createdclusterrolebinding.rbac.authorization.k8s.io/metrics-server:system:auth-delegator createdclusterrolebinding.rbac.authorization.k8s.io/system:metrics-server createdservice/metrics-server createddeployment.apps/metrics-server createdapiservice.apiregistration.k8s.io/v1beta1.metrics.k8s.io created

4、查看

[root@k8s-master ~]# kubectl get pod -ANAMESPACE NAME READY STATUS RESTARTS AGEkube-system calico-kube-controllers-658d97c59c-qjdrt 1/1 Running 12 (56m ago) 6h48mkube-system calico-node-5msx4 1/1 Running 0 6h48mkube-system calico-node-5q749 1/1 Running 2 (54m ago) 6h48mkube-system calico-node-z9gtq 1/1 Running 0 6h48mkube-system coredns-66f779496c-2v6l9 1/1 Running 1 7h18mkube-system coredns-66f779496c-kj6db 1/1 Running 3 (81m ago) 7h18mkube-system etcd-k8s-master 1/1 Running 7 7h18mkube-system kube-apiserver-k8s-master 1/1 Running 6 7h18mkube-system kube-controller-manager-k8s-master 1/1 Running 6 (54m ago) 7h18mkube-system kube-proxy-77vqw 1/1 Running 0 7h18mkube-system kube-proxy-cmnxt 1/1 Running 0 7h16mkube-system kube-proxy-ksrmj 1/1 Running 0 7h17mkube-system kube-scheduler-k8s-master 1/1 Running 6 (55m ago) 7h18mkube-system metrics-server-794bb5d868-ghk49 1/1 Running 0 61s

可见已在Kube-system名称空间中成功运行

5、测试

[root@k8s-master ~]# kubectl top nodeNAME CPU(cores) CPU% MEMORY(bytes) MEMORY% k8s-master 199m 9% 1520Mi 41% k8s-node1 77m 3% 914Mi 53% k8s-node2 87m 4% 911Mi 52%
[root@k8s-master ~]# kubectl top pod -ANAMESPACE NAME CPU(cores) MEMORY(bytes) kube-system calico-kube-controllers-658d97c59c-qjdrt 1m 11Mi kube-system calico-node-5msx4 45m 105Mi kube-system calico-node-5q749 44m 69Mi kube-system calico-node-z9gtq 42m 128Mi kube-system coredns-66f779496c-2v6l9 2m 13Mi kube-system coredns-66f779496c-kj6db 2m 11Mi kube-system etcd-k8s-master 25m 86Mi kube-system kube-apiserver-k8s-master 76m 288Mi kube-system kube-controller-manager-k8s-master 23m 50Mi kube-system kube-proxy-77vqw 1m 18Mi kube-system kube-proxy-cmnxt 1m 31Mi kube-system kube-proxy-ksrmj 1m 36Mi kube-system kube-scheduler-k8s-master 3m 22Mi kube-system metrics-server-794bb5d868-ghk49 4m 15Mi