In the linked BF, we are seeing an unexpectedly high planning time metric in slow query logs. Based on the query and indexes involved, I don't expect the planner itself to be taking so long. Instead, it seems to me that the planner is waiting for something unrelated to planning, but it's unclear what. If so, it's possible that SERVER-92476 didn't solve the whole problem, and there is still some non-planning code being included in the planning time metric.

To help narrow down where the wait is happening, I'm going to increase the log verbosity in the test, then revisit the BF if it happens again.