Explore how we can get metrics when connecting to spoke clusters. There are at least two approaches:

1. Query metrics from the hub cluster. Spokes send metrics to the hub.

• We need to determine if there are any gaps here. Currently this will not let us get alerts from the spokes.
• Need to determine how RBAC works. Will normal users be able to get metrics?

2. Query metrics directly from the public monitoring endpoints on each spoke cluster.

• Need to determine how to discover the monitoring URLs. The monitoring endpoints are available in the monitoring-shared-config config map in the openshift-config-managed namespace on each spoke. ACM might need to add this to the managed cluster resource.
• Currently the tenancy port is not exposed outside the cluster through a route. This would create a dependency on the monitoring team to expose this service port for normal users to see metrics.

Acceptance Criteria

• Determine which of the two approaches is feasible.
• Explore how metrics work in ACM and understand any limitations using the hub for metrics.
• Reach out to the monitoring team to get their input.
• Document findings and recommend an approach for implementation.

Non Goals

This story does not cover any implementation, only R&D.