tencent cloud

14天试用边缘安全加速平台 EO 限时免费

Feedback

Tencent Kubernetes Engine

Installing metrics-server on TKE

Last updated: 2024-12-13 21:25:13

Operation Scenarios

The metrics-server can realize the Resource Metrics API (metrics.k8s.io) of Kubernetes. Through this API, you can query some monitoring metrics of Pods and Nodes. The monitoring metrics of Pods are used in HPA, VPA, and kubectl top pods commands, whereas the Node metrics are currently used only in kubectl top nodes commands. TKE itself realizes the Resource Metrics API, pointed towards the hpa-metrics-server, and currently TKE also provides monitoring metrics for Pods.
After installing the metrics-server to the cluster, you can run kubectl top nodes to obtain the monitoring overview of nodes to replace the realization of the Resource Metrics API. HPA created on the TKE console does not use Resource Metrics and only uses Custom Metrics. Therefore, installing the metrics-server does not affect HPA created on the TKE console. This document describes how to install the metrics-server on TKE.

Directions

Downloading the YAML deployment file

Run the following commands to download the latest deployment file components.yaml of the metrics-server.
wget https://github.com/kubernetes-sigs/metrics-server/releases/latest/download/components.yaml

Modifying the metrics-server launch parameter

The metrics-server requests the kubelet API of each node to obtain monitoring data. The API is exposed via HTTPS, but as TKE node kubelet uses a self-signed certificate, if the metrics-server directly requests the kubelet API, an error of certification verification failure will occur. Therefore, you need to add the --kubelet-insecure-tls launch parameter in the components.yaml file. Moreover, as the official image repository of the metrics-server is stored in k8s.gcr.io, users in China may not be able to directly pull images from the repository. You need to manually synchronize images to CCR or use the synchronized image ccr.ccs.tencentyun.com/mirrors/metrics-server:v0.4.0.
Below is a sample of modification of the components.yaml file:
containers:
- args:
- --cert-dir=/tmp
- --secure-port=4443 # Please replace with 4443
- --kubelet-preferred-address-types=InternalIP,ExternalIP,Hostname
- --kubelet-use-node-status-port
- --kubelet-insecure-tls # Add this launch parameter
image: ccr.ccs.tencentyun.com/mirrors/metrics-server:v0.4.0 # For cluster in the Chinese mainland, please replace with this image address
ports:
- containerPort: 4443 # Please replace with 4443
name: https
protocol: TCP

Deploying the metrics-server

After modifying components.yaml, run the following commands to implement one-click deployment to the cluster via kubectl:
kubectl apply -f components.yaml
Note:
Through the above step, you can install and deploy the metrics-server. Alternatively, you can run the following commands for one-click installation of the metrics-server, but this method cannot ensure synchronization with the latest version.
kubectl apply -f https://raw.githubusercontent.com/TencentCloudContainerTeam/manifest/master/metrics-server/components.yaml

Checking the running status

1. Run the following commands to check whether the metrics-server starts normally. Below is a sample:
$ kubectl get pod -n kube-system | grep metrics-server
metrics-server-f976cb7d-8hssz 1/1 Running 0 1m
2. Run the following commands to check the configuration file. Below is a sample:
$ kubectl get --raw /apis/metrics.k8s.io/v1beta1 | jq
{
"kind": "APIResourceList",
"apiVersion": "v1",
"groupVersion": "metrics.k8s.io/v1beta1",
"resources": [
{
"name": "nodes",
"singularName": "",
"namespaced": false,
"kind": "NodeMetrics",
"verbs": [
"get",
"list"
]
},
{
"name": "pods",
"singularName": "",
"namespaced": true,
"kind": "PodMetrics",
"verbs": [
"get",
"list"
]
}
]
}
3. Run the following commands to check the node usage performance. Below is a sample:
$ kubectl top nodes
NAME CPU(cores) CPU% MEMORY(bytes) MEMORY%
test1 1382m 35% 2943Mi 44%
test2 397m 10% 3316Mi 49%
test3 81m 8% 464Mi 77%
Contact Us

Contact our sales team or business advisors to help your business.

Technical Support

Open a ticket if you're looking for further assistance. Our Ticket is 7x24 avaliable.

7x24 Phone Support
中国香港
+852 800 906 020 (免费)
美国
+1 844 606 0804 (免费)
英国
+44 808 196 4551 (免费)
加拿大
+1 888 605 7930 (免费)
澳大利亚
+61 1300 986 386 (免费)
EdgeOne 热线
+852 300 80699
更多本地服务热线陆续新增中