On Mon, 6 Oct 2025 at 23:39, Marko Tiikkaja <marko@joh.to> wrote:
> On the production server the query still consistently took ~100ms even
> after all that maintenance had been done. Compared to <0.05ms when
> going through the primary key.
This part surprises me. Is this on a standby server or primary? Are
you certain that no rows are matching the state = 'WAIT_EVENT' qual?
David