Monitor shared runner usage per project

As per title, is this possible at all?

I was hoping to be able to determine which projects were using the most shared runner resources to help the teams optimised, and to determine when we may need to expand the cluster (I’m using Kubernetes).

None of the GitLab Runner metrics seem to expose project information, so I don’t see how I can determine which project is burning through all the cycles.

Am I simply looking in the wrong place? Or is there some magic I need to do in the Helm values.yaml to make this possible?

Maybe an alternative approach: On GitLab.com, shared runner usage is calculated with compute minutes, and quotas. The same features exist on self-managed instances with instance runners but are disabled by default. Compute minutes administration | GitLab I heard of a customer setup where the platform team uses compute minutes to provide internal budgets for different teams, and also for scaling insights.