On Thu, Feb 03, 2022 at 07:26:01PM -0800, Andres Freund wrote: > Which reminds me: Perhaps we ought to hint about reducing / removing > autovacuum cost limits in this situation? And perhaps make autovacuum absorb > config changes while running? It's annoying that an autovac halfway into a > huge table doesn't absorb changed cost limits for example.
I remembered this thread: https://commitfest.postgresql.org/32/2983/ | Running autovacuum dynamic update to cost_limit and delay https://www.postgresql.org/message-id/flat/13A6B954-5C21-4E60-BC06-751C8EA469A0%40amazon.com https://www.postgresql.org/message-id/flat/0A3F8A3C-4328-4A4B-80CF-14CEBE0B695D%40amazon.com -- Justin
