Comment by claytonjy

This is intriguing; could you give an example of a distribution Postgres struggles with at scale, and an application that produces such a distribution?