We have observed occasional issues in the past where attempting to decode table statistics histograms causes a panic (e.g., https://github.com/cockroachdb/cockroach/issues/128876). We catch panics in some parts of the stats cache, but there are some paths through the stats cache where panics are not caught. This causes certain operations that access the stats cache to crash the cockroach process. We should add panic catching for these alternate paths.
We have observed occasional issues in the past where attempting to decode table statistics histograms causes a panic (e.g., https://github.com/cockroachdb/cockroach/issues/128876). We catch panics in some parts of the stats cache, but there are some paths through the stats cache where panics are not caught. This causes certain operations that access the stats cache to crash the cockroach process. We should add panic catching for these alternate paths.
Jira issue: CRDB-44798