Metrics Server is added as a tech preview feature for 4.15. We need to add content to the Monitoring overview topic to account for this new component, noting where needed that it is TP only for this release.

Details of metrics server implementation

• Replace CMO's prometheus-adapter deployment with metrics-server as the provider for the metrics.k8s.io APIService
• OpenShift gets an operator that deploys and reconciles metrics-server and related resources

Why is this change important?

• For autoscaling, OpenShift needs a resource metrics implementation. Currently, this depends on CMO's Prometheus stack
• Some users would like to opt-out of running a full-fledged Prometheus stack: see https://issues.redhat.com/browse/MON-3152

Scenarios

1. If a cluster admin decides to not deploy a full Monitoring stack, autoscaling must still work.

Other info

• https://kubernetes.io/docs/tasks/debug/debug-cluster/resource-metrics-pipeline/#metrics-server
• https://github.com/openshift/kubernetes-metrics-server