only keep some of the metrics in the object storage

[yasharne]

Is your proposal related to a problem?

Most of the metrics we are collecting in our Prometheus server does not worth keeping for long, like CPU usage and disk usage. on the other hand, some other metrics are very important like the ones that describe the business status so that's why we wanted to use Thanos to keep those important metrics longer

Describe the solution you'd like

Add the ability to define which metrics I want to keep longer and which ones I don't want to keep longer so it wouldn't occupy the object storage with the useless data

[GiedriusS]

I believe this feature request is covered by: #903

[GiedriusS]

I assume that the thumbs up means agreement so closing this to avoid duplicates 👍