Description
We're seeing the number of regions the stats table thinks there are not being in agreement with the actual number of regions. It seems to be related to splits and compactions. In particular, when a split is occurring, the deletion of the stats for the region being split is not being deleted as expected.