I believe your autovacuum_vacuum_insert_threshold regression will *mostly* go away following this bug fix: git.postgresql.org/gitweb/?p=post -- thanks again!
Conversation
Replying to
I look forward to my next round of benchmarks with Postgres and I appreciate that it continues to improve.
1
5
Replying to
May now make sense to set autovacuum_vacuum_insert_threshold very low with a 100% append-only table, even on 13. Postgres 14 will be able to do the same thing with ~99% append-only table, thanks to a pending patch by . (I recall you asked about this separately.)

