Re: n_ins_since_vacuum stats for aborted transactions

On Wednesday, April 9, 2025, Sami Imseih <samimseih(at)gmail(dot)com> wrote:

>
> > What is the use case for that behavior? Perhaps you have one, but until
> you make it explicit, it is hard for others to get behind your proposal.
>
> The point is to ensure that the pg_stats fields that autovacuum uses
> are supplied the correct values
> for the different thresholds they need to calculate, as described here [0]
>
>
> [0] https://www.postgresql.org/message-id/CAA5RZ0uDyGW1omWqWkxyW8NB1qzsK
> mXhnoMtzTBeRzSd4DMatQ%40mail.gmail.com
>
>
Except there isn’t some singular provably correct value here. Today’s
behavior (considering dead tuples) is not intrinsically wrong nor correct,
and neither is what you propose (ignoring the dead tuples). The fact that
those dead tuples get removed regardless is a point in favor of counting
them when deciding what to do. And it’s also the long-standing behavior.
You need to make a compelling argument to change to your preference.

Inserting aborted dead tuples moves the counter closer to both autovacuum
thresholds. There is no reason why that should be prohibited. I can see
the argument for why one threshold should be dead tuples only and the other
live tuples only - but I don’t favor that design.

David J.