[spike] improve must-gather for Monitoring

Currently we don't leverage a lot the must-gather infrastructure for debugging lower level issues. We have limited coarse grained telemetry data, but a reoccurring theme is to inspect the following:

• A prometheus TSDB dump. This is potentially quite big in size. Maybe we could cap it somehow.
• Get a snapshot of the state for alertmanagers, targets, config, runtimeinfo as per https://bugzilla.redhat.com/show_bug.cgi?id=1845561#c18

DoD:

• improve must-gather debugging insights for Monitoring