[common][rabbitmq] optionally add detailed metrics

[Carthaca]

This enables https://www.rabbitmq.com/docs/prometheus#detailed-endpoint via podmonitor (not servicemonitor, because the metrics port is not present at the service).

E.g. gives insights into the queue names via parameter family=queue_coarse_metrics in new metrics called rabbitmq_detailed_XXX which are added on top.

[bashar-alkhateeb]

It is part of the task we are working on it https://jira.tools.sap/browse/CCC-34487 https://github.com/sapcc/helm-charts/pull/6033 but we need to add more endpoint @djdule would you please check this PR and let us discuss which approach we need to use , using a subchart or creating the podmonitor resource in rabbtimq helm chart ?

[bashar-alkhateeb]

@Carthaca would you check PR: https://github.com/sapcc/helm-charts/pull/6033 if that provides the same needed metric please close this PR

[Carthaca]

The params are mandatory for the detailed endpoint, see https://www.rabbitmq.com/docs/prometheus#detailed-endpoint

As mentioned earlier, using per-object metrics in environments with a lot of entitites is very computationally expensive. For example, /metrics/per-object returns all metrics for all entities in the system, even if many of them are not used by most clients (such as monitoring tools).

This is why there is a separate endpoint for per-object metrics that allows the caller to query only the metrics they need:

GET /metrics/detailed

By default it does not return any metrics. All required metric groups and virtual host filters must be be provided as query parameters.

[bashar-alkhateeb]

Oops! .. I think we forgot that in the servicemonitor resource we will try to add this to the servicemonitor soon, as well as the related Alert.