Feature: expose more Prometheus metrics than just /minio/v2/metrics/cluster

[BZValoche]

I need to have bucket metrics in grafana. I know that I can create a ScrapeConfig for it, but it is frustrating because there is some manual actions in order to achieve this, and my environment is completely managed by Argocd.

When you set prometheusOperator: true in a Tenant values, it create a scape config only on /minio/v2/metrics/cluster, which is hardcoded in the constant MinIOPrometheusPathCluster in constants.go.
Minio provides the following v2 metrics paths:

• cluster
• bucket
• node
• resource

It would be nice to have the possibility to enable the metrics or not (the values should mention the PROMETHEUS_NAMESPACE which has to be set at the operator level, btw).
A possible implementation would be to add a configuration option which could look like this:

prometheusOperatorScrape:
  cluster: true
  bucket: true
  node: false
  resource: false

Or maybe a list, so if minio adds metrics there is no need to modify the controller:

prometheusOperatorScrapes:
- cluster
- bucket

As a final word, the ScrapeInterval is also hardcoded, it'd also be nice to have it configurable.

If you find this of interest, I am happy to help with a PR.

[lm-solution]

for me it was confusing, that the operator uses the additionalScrapeConfig to ingest the scraping drect into prometheus.
wouldnt it be much nicer to use a servicemonitor CR.
this would be best practice and much more transparent and less invasive for the deployment.