Ceph » mgr

Issue:
currently ceph mgr exports it's metrics in Prometheus format making them unable to be parsed and handled by opentelemtry.
This leaves companies with a large-scale agreement on opentelemetry scraping in a position to either write a manual scraper, handling exceptions of clashing name and type conversion,
or drop monitoring of ceph mgr metric values at all.

Suggestion:
providing a possibility to choose between the metric type being exported by ceph mgr.
With containerized deployments it should also be possible to use opentelemetry aside the currently expose prometheus metrics in addition.

Business justification:
The clusters we are trying to monitor and integrate into our global "opentelemetry policy" are in total ~64PB serving more than 1500+ direct clients connecting to the clusters.