On a SNO spoke, MCOA collects 4k metrics while the metrics-collector only collects 2.5k metrics. Some clients are highly sensitive the the amount of resources used by our stack and more generally red-hat. We must ensure that we collect only what is necessary.

Acceptance:

• Identify the additional metrics source compared to current metrics collector and remove them (it might simply be the dynamic ones that are only collected based on a trigger).
• As a permanent solution, we introduce a metrics list generator that extracts the metrics and labels needed for the dashboards. This list is then consumed for generating the default list of metrics for platform monitoring. CI fails if the list is not up to date?
• Ensure that the defined rules are valid. Some are not computing anything and thus probably not needed.