phlare

phlare copied to clipboard

Right now we're initialisation all metrics in the new of PhlareDB which means if something fails after, we'll never be able to re-register metrics again.

This means we need to refactor our metrics and dependencies in PhlareDB to pass them down instead of registering them again. At the same time, we should get rid of tenant metrics for most of our metrics, if we still want per-tenant metrics then the label should be added during metric observation.

see https://github.com/grafana/phlare/pull/340