Feedback-ID: i0fe9450f:Fastmail
Date: Wed, 25 Mar 2026 12:16:12 +0900
From: Michael Paquier <michael@paquier.xyz>
To: Sami Imseih <samimseih@gmail.com>
Cc: Bertrand Drouvot <bertranddrouvot.pg@gmail.com>,
	Fujii Masao <masao.fujii@gmail.com>,
	pgsql-hackers@lists.postgresql.org,
	Zsolt Parragi <zsolt.parragi@percona.com>
Subject: Re: Flush some statistics within running transactions
Message-ID: <acNTfL1xO_UUXkZQ@paquier.xyz>
References: 
 <CAA5RZ0u2Vi4-PvHsFBS6aHpoi-uNQjDaLxnAwXxRuUf6ZDMs3Q@mail.gmail.com>
 <aZiD74ZVehxfgqxM@ip-10-97-1-34.eu-west-3.compute.internal>
 <CAA5RZ0t1DsB5x_reGAv0AcKdKuF5FTowUx54SLnWkD3w5vH4Lg@mail.gmail.com>
 <aZwMdVBYobV41duU@ip-10-97-1-34.eu-west-3.compute.internal>
 <CAA5RZ0vaB+fXHt76bi_Koj33JoerdgfMYFxHn-D1N+SPSGwL2w@mail.gmail.com>
 <aZ2TGo5bkDvYOfIw@ip-10-97-1-34.eu-west-3.compute.internal>
 <abeimWKYq-n34FrH@paquier.xyz>
 <abfLaVT5qIbPtzsQ@ip-10-97-1-34.eu-west-3.compute.internal>
 <abfZ55jwCGGci6aK@paquier.xyz>
 <CAA5RZ0ssPraeF9gg8S8G7O25t90t8_Y2PU-jdNkZ42EiNH22CA@mail.gmail.com>
MIME-Version: 1.0
Content-Type: multipart/signed; micalg=pgp-sha512;
	protocol="application/pgp-signature"; boundary="URaJ+n8gRt1tPalr"
Content-Disposition: inline
In-Reply-To: 
 <CAA5RZ0ssPraeF9gg8S8G7O25t90t8_Y2PU-jdNkZ42EiNH22CA@mail.gmail.com>
Archived-At: 
 <https://www.postgresql.org/message-id/acNTfL1xO_UUXkZQ%40paquier.xyz>
Precedence: bulk


--URaJ+n8gRt1tPalr
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline

On Mon, Mar 16, 2026 at 06:42:44PM -0500, Sami Imseih wrote:
> So attached is a new proposal with tests and docs. In terms of
> test I fell back to the strategy used by Bertrand [0] with the

As far as I am reading the patch, it seems to me that this would
accelerate the frequency of the stats flushes when we have a
transaction with many short queries, but it does not help much in the
case of an analytical single query because the flush of the stats
would just happen once after the timeout set for each query?

Let's imagine for example a query that takes hours to run, where we'd
want to get fresh IO and WAL stats (backend included) on a periodic
basis by processing interrupts while going through the executor.

Thinking more about this problem, I'd really like to think that a
client-side API would provide a more flexible interface for the
analytical case.
--
Michael

--URaJ+n8gRt1tPalr
Content-Type: application/pgp-signature; name=signature.asc

-----BEGIN PGP SIGNATURE-----
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=QiZW
-----END PGP SIGNATURE-----

--URaJ+n8gRt1tPalr--