1. Proposed title of this feature request

Expose Karpenter/AutoNode metrics to ROSA/ARO HCP cluster data plane

2. What is the nature and description of the request?

As a ROSA HCP cluster administrator using Karpenter (AutoNode), I need access to Karpenter metrics from my cluster's monitoring stack (Prometheus/CMO) so I can:

• Monitor Karpenter's provisioning decisions and behavior
• Track node scaling activities and trends
• Set up alerts based on Karpenter operational metrics
• Integrate Karpenter metrics with existing observability pipelines

Currently, Karpenter runs on the management cluster (HCP namespace) and its metrics are not exposed to the hosted cluster's data plane. Only apiserver metrics are accessible because that endpoint is visible from the data plane.

3. Why does the customer need this? (List the business requirements here)

Without visibility into Karpenter metrics, customers cannot effectively monitor their node provisioning, troubleshoot scaling issues, or perform capacity planning

4. List any affected packages or components.

• ROSA/ARO HyperShift / Hosted Control Planes
• Karpenter controller
• Cluster Monitoring Operator (CMO)
• Platform monitoring stack on hosted clusters

Per the HyperShift team: An additional endpoint on the control plane side would aggregate and expose all metrics endpoints from control plane workloads. The platform monitoring stack would then scrape that endpoint from inside the hosted cluster.