Hi,
On Tue, Feb 03, 2026 at 08:37:31PM -0600, Sami Imseih wrote:
> > > I wonder if we can use GetCurrentStatementStartTimestamp()
> > > instead?
> >
> > The main use case for this column is being able to retrieve the stats
> > that were updated since the last time one was looking. That only works
> > if it's the statement end time, or else long-running statements
> > spanning more than one poll interval would be missed.
>
> Sure, I get it is not perfect for the polling use-case due to the scenario
> you mention, but I don't think it will be acceptable to call
> GetCurrentTimeStamp() at the end of every execution and especially
> with a SpinLock held.
I think the same, that would not match (lmgr/README):
"
* Spinlocks. These are intended for *very* short-term locks. If a lock
is to be held more than a few dozen instructions, or across any sort of
kernel call
"
Out of curiosity I looked for GetCurrentTimeStamp() calls while holding a spinlock
and found one in WalReceiverMain(). But I guess it's less of an issue since
it's only called when the walreceiver starts.
Regards,
--
Bertrand Drouvot
PostgreSQL Contributors Team
RDS Open Source Databases
Amazon Web Services: https://aws.amazon.com
