参考//metrics-server部署 - 简书
下载yaml文件和镜像
wget https://github.com/kubernetes-sigs/metrics-server/releases/latest/download/components.yaml
这个yaml文件中的metrics-server镜像地址为外国地址，由于墙的原因下不了，需要先手动到Docker Hubhttps://hub.docker.com/
下载，我这里下载的版本是V0.5.0
手动换一下标签
sudo docker pull bitnami/metrics-server:0.5.0 sudo docker tag bitnami/metrics-server:0.5.0 k8s.gcr.io/metrics-server/metrics-server:v0.5.0 sudo docker rmi bitnami/metrics-server:0.5.0
然后修改yaml文件
在args：后增加 --kubelet-insecure-tls -不要验证Kubelets提供的服务证书的CA。 将其部署到master节点上，不然镜像需要在每一个分配的节点都下载一遍，很麻烦 给master节点打标签 metrics-server="true" 在deployment.spec.template.spec下增加 nodeSelector: metrics-server: "true" tolerations: - key: "node-role.kubernetes.io/master" operator: "Exists" effect: "NoSchedule"
然后部署yaml
执行kubectl top nodes查看节点资源使用情况
NAME CPU(cores) CPU% MEMORY(bytes) MEMORY% k8s70131 336m 8% 3328Mi 42% k8s70132 338m 8% 3569Mi 45% k8s70133 356m 8% 3760Mi 47% k8s70134 202m 10% 1707Mi 44% k8s70135 211m 1% 5615Mi 36% k8s70136 84m 1% 3908Mi 16% k8s70137 164m 4% 2424Mi 30%
执行kubectl top pod -n harbor 查看某个命名空间pod资源使用情况
NAME CPU(cores) MEMORY(bytes) harbor-chartmuseum-67db84bb8c-s4wgd 1m 15Mi harbor-core-5bbbc85447-qxpnr 2m 43Mi harbor-database-0 10m 61Mi harbor-jobservice-7c9ff58f7-xqcdp 2m 14Mi harbor-notary-server-b8c5ff4b8-kl6jt 1m 7Mi harbor-notary-signer-847466489-pqdxd 3m 7Mi harbor-portal-c55c48545-zrwsq 1m 3Mi harbor-redis-0 3m 3Mi harbor-registry-84f6454ff8-bd6t4 1m 17Mi harbor-trivy-0 2m 8Mi
