Hmm, let me try my question another way. What is the expected way to monitor the Rancher services, components and cluster connections? I'm not referring to installing prometheus in the downstream clusters to monitor those clusters, but to monitor the rancher installation itself. And if this generally isn't done, as evidenced by the old prometheus exporter which is now archived, why not? What is generally done instead?

Not sure if this answers your question but maybe it helps: https://ranchermanager.docs.rancher.com/integrations-in-rancher/monitoring-and-alerting/how-monitoring-works#default-exporters