22 lines
720 B
Markdown
22 lines
720 B
Markdown
# Metrics Server
|
|
|
|
Metrics Server provides the Kubernetes Metrics API for live resource usage,
|
|
including `kubectl top` and future HPA/VPA resource metrics.
|
|
|
|
It is intentionally small and not scraped by Prometheus. Prometheus/Grafana
|
|
remain the source for historical dashboards and alerting.
|
|
|
|
Talos kubelets use serving certificates that are not trusted by metrics-server
|
|
unless kubelet server certificate rotation and CSR approval are configured.
|
|
This deployment uses `--kubelet-insecure-tls` so it works with the current
|
|
Talos configuration.
|
|
|
|
After syncing, verify:
|
|
|
|
```bash
|
|
kubectl -n kube-system get deploy metrics-server
|
|
kubectl get apiservice v1beta1.metrics.k8s.io
|
|
kubectl top nodes
|
|
kubectl top pods -n litellm
|
|
```
|