When we find an eligible time-series bucket on-disk to reopen for an insert, we perform some checks before bringing it back into memory in the bucket catalog here. At the moment, we increment a counter any time that one of these checks fail, but all of these different failures increment the same counter.

It could be useful to additionally have some finer granularity counters around specific reasons that the checks failed (e.g. due to era mismatch, hash collision, or malformed bucket).

This is a spin-off from SERVER-98672.