已应用Kubernetes Dashboard YAML却无服务/部署，如何用公网IP暴露服务？

It sounds like you’ve hit a common snag when setting up the Kubernetes Dashboard—let’s break down the likely issues and how to fix them:

1. Verify the YAML was actually applied successfully #

First things first: did the kubectl apply command complete without errors? It’s easy to miss a typo in the command or YAML path that causes resource creation to fail.

• Re-run the apply command and check the output closely:

  sudo kubectl apply -f <your-dashboard-yaml-file>
  

  You should see confirmation messages like deployment.apps/kubernetes-dashboard created or service/kubernetes-dashboard created. If you get syntax errors, permission warnings, or path not found messages, that’s your first fix.

• Check recent cluster events for hidden failures:

  sudo kubectl get events --sort-by='.metadata.creationTimestamp'
  

  This will show if there were issues pulling the Dashboard image, scheduling pods, or creating services that didn’t show up in the apply command output.

2. Check the correct namespace #

The official Kubernetes Dashboard is deployed to the kube-system namespace by default—but when you run kubectl get services or kubectl get deployment without specifying a namespace, it only checks the default namespace.

• List all resources in the kube-system namespace to confirm Dashboard exists:

  sudo kubectl get all -n kube-system | grep dashboard
  

  If you see deployment, service, and pod entries here, that means the resources are live—you just weren’t looking in the right place!

3. Validate your Dashboard YAML file #

If the apply command succeeded but resources still don’t show up, double-check your YAML content:

• Ensure the apiVersion values match your Kubernetes cluster version (e.g., apps/v1 for deployments is required in K8s 1.19+).
• Confirm the kind fields are correctly set to Deployment and Service for the core Dashboard resources.
• If you modified the YAML to use a custom namespace, make sure the metadata.namespace field is set properly, and adjust your kubectl commands to target that namespace.

4. Check cluster and kubectl context health #

• Make sure all your cluster nodes are in a healthy state:

  sudo kubectl get nodes
  

  Any nodes marked NotReady could prevent Dashboard pods from being scheduled.
• Verify you’re connected to the correct cluster:

  sudo kubectl config current-context
  

  If you have multiple clusters configured, it’s possible you applied the YAML to a different cluster than the one you’re checking.

Once you locate the missing resources, you can proceed with exposing the service using a NodePort or LoadBalancer type (depending on your cluster setup) to access it via your server’s public IP.

内容的提问来源于stack exchange，提问作者Warok