Auto-Scaling === ###### tags: `III` `DevOps` `Infra` ## Terminology - 水平擴展 scale out: 增加服務的 pod、增加服務的機器 - 垂直擴展 scale up: 同一台服務的機器增加 cpu、memory ## Add worker node 在 master node ``` kubeadm token list ``` ``` kubeadm token create --print-join-command ``` 在 worker node ``` kubeadm join 192.168.2.16:6443 \ --token s8o9wi.dylbvs735sy53mmq \ --discovery-token-ca-cert-hash sha256:0c16a05978533ca8f44af6e779162a1c99516fa2a4acd81915f0379755a856bc ``` 會用到的 [kubeadm join 的參數]( https://blog.scottlowe.org/2019/08/15/reconstructing-the-join-command-for-kubeadm/) --- ## Useful Commands - [K8s 安裝筆記 - kubeadm 手動 (ubuntu 16.04 / ubuntu 18.04)](https://rickhw.github.io/2019/03/17/Container/Install-K8s-with-Kubeadm/) - [kubeadm reset](https://kubernetes.io/docs/reference/setup-tools/kubeadm/kubeadm-reset/) - [reset 參考命令](http://gitvepc/abby/tools/-/blob/master/k8s_init_setup_bash.sh) --- ## Metrics Server HPA 會用到 metrics server 提供的 metrics * Only for CPU & memory ``` helm install -f metric-server-values.yml --namespace kube-system gitlab-runner stable/metrics-server ``` ``` # metric-server-values.yml args: ["--kubelet-insecure-tls", "--kubelet-preferred-address-types=InternalIP"] ``` 等一下，測試 ``` kubectl get --raw "/apis/metrics.k8s.io/v1beta1/nodes" kubectl top nodes kubectl get --raw "/apis/metrics.k8s.io/v1beta1/pods" kubectl top pod -n <NAMESPACE> ``` * [Setting up a Custom Metrics Server](https://github.com/stefanprodan/k8s-prom-hpa#setting-up-a-custom-metrics-server) * Prometheus http://10.22.100.231:31190/ http://10.22.100.233:31190/ * Custom Metrics Server ``` kubectl get --raw "/apis/custom.metrics.k8s.io/v1beta1/namespaces/5gc/pods/*/gitlab_runner_version_info" | jq ```