Metrics Server installation in Kubernetes - Step-by-Step Execution

Process Flow - Metrics Server installation

Download Metrics Server YAML

↓

Apply YAML to cluster

↓

Metrics Server Pods start

↓

Check Pod status

↓

Metrics Server running

↓

kubectl top nodes/pods works

This flow shows downloading the Metrics Server setup file, applying it to the cluster, starting pods, checking their status, and verifying metrics availability.

Execution Sample

Kubernetes

kubectl apply -f https://github.com/kubernetes-sigs/metrics-server/releases/latest/download/components.yaml
kubectl get pods -n kube-system
kubectl top nodes

These commands install Metrics Server, check pod status, and display node metrics.

Process Table

Step	Command	Action	Result/Output
1	kubectl apply -f https://github.com/kubernetes-sigs/metrics-server/releases/latest/download/components.yaml	Apply Metrics Server manifests	deployment.apps/metrics-server created serviceaccount/metrics-server created clusterrole.rbac.authorization.k8s.io/metrics-server created ...
2	kubectl get pods -n kube-system	Check Metrics Server pod status	NAME READY STATUS RESTARTS AGE metrics-server-abcdef123-xyz 1/1 Running 0 30s
3	kubectl top nodes	Query node metrics	NAME CPU(cores) MEMORY(bytes) worker-node1 100m 200Mi
4	kubectl top pods	Query pod metrics	NAME CPU(cores) MEMORY(bytes) metrics-server-abcdef123-xyz 5m 20Mi
5	kubectl get pods -n kube-system	If pod not ready	NAME READY STATUS RESTARTS AGE metrics-server-abcdef123-xyz 0/1 CrashLoopBackOff 3 2m
6	Troubleshoot logs	Check pod logs	Error: unable to fetch metrics from Kubelet
7	Fix args in deployment	Add --kubelet-insecure-tls flag	Pod restarts and becomes Ready
8	kubectl top nodes	Verify metrics after fix	NAME CPU(cores) MEMORY(bytes) worker-node1 100m 200Mi

💡 Metrics Server pod is running and metrics commands return data successfully

Status Tracker

Variable	Start	After Step 1	After Step 2	After Step 3	After Step 7	Final
Metrics Server Pod Status	Not created	Creating	Running (1/1)	Running (1/1)	Restarting with fix	Running (1/1)
Metrics Availability	No metrics	No metrics	No metrics	No metrics	Metrics shown	Metrics shown

Key Moments - 3 Insights

Why does 'kubectl top nodes' show no metrics right after installation?

Because the Metrics Server pod may not be fully running yet (see step 2 in execution_table), so metrics are not available until pod is Ready.

What does it mean if the Metrics Server pod is in CrashLoopBackOff?

It means the pod is repeatedly failing to start, often due to TLS or permission issues. See step 5 and 6 for troubleshooting logs.

Why add the --kubelet-insecure-tls flag in the deployment?

This flag allows Metrics Server to connect to kubelets without strict TLS verification, fixing common connection errors shown in step 6.

Visual Quiz - 3 Questions

Look at the execution_table, at which step does the Metrics Server pod first show as Running?

A Step 2

B Step 1

C Step 5

D Step 7

💡 Hint

Check the 'Result/Output' column for pod status in step 2.

According to variable_tracker, what is the Metrics Availability state after Step 3?

A Metrics shown

B Pod not created

C No metrics

D Pod crashing

💡 Hint

Look at the 'Metrics Availability' row under 'After Step 3' column.

If the Metrics Server pod is in CrashLoopBackOff, what should you do next according to the execution_table?

A Run 'kubectl top nodes' again

B Check pod logs and fix deployment args

C Delete the pod and reinstall Kubernetes

D Ignore and wait

💡 Hint

See steps 5 and 6 for pod status and logs, and step 7 for the fix.

Concept Snapshot

Metrics Server installation:
1. Download and apply Metrics Server YAML.
2. Check pod status in kube-system namespace.
3. Use 'kubectl top' to view metrics.
4. Troubleshoot pod issues with logs.
5. Add --kubelet-insecure-tls if TLS errors occur.
6. Confirm metrics availability after pod is Ready.

Full Transcript

To install Metrics Server in Kubernetes, first download and apply the official YAML manifest. This creates the Metrics Server deployment and related resources. Next, check the pod status in the kube-system namespace to ensure the pod is running. If the pod is not ready or crashes, check the pod logs for errors such as TLS connection issues. A common fix is to add the --kubelet-insecure-tls flag to the deployment arguments to allow insecure TLS connections to kubelets. Once the pod is running, use 'kubectl top nodes' and 'kubectl top pods' to verify metrics are available. This process ensures Metrics Server is properly installed and functional.