On Fri, Apr 11, 2025 at 3:47 AM Robert Haas <robertmhaas@gmail.com> wrote: > I'd probably write the increments as ++ rather than += 1 but I'm not > sure if everyone would agree.
I'm a fan of ++ as well, so I agree.
Thanks Richard
Looking at this patch and the idea altogether I have other thoughts on this now.
Thinking about the original problem of having a count of total idle in transaction time, seems unnecessary with the availability of state_change timestamp along with state columns in pg_stat_activity.