Update: this commit didn’t change anything. Database still roaring at >95% constantly after update and killing all old running queries. At beta4 it used to run at a 20-30% stable cadence for this forum. We already ran an auto-vacuum and schema reindex as well.
How do we disable this feature? It doesn’t seem right to get a larger database instance to resolve this, there seems to be something wrong with how this query is computed based on the fact it ran great before this update.