| From: | Ron Johnson <ronljohnsonjr(at)gmail(dot)com> |
|---|---|
| To: | Pgsql-admin <pgsql-admin(at)lists(dot)postgresql(dot)org> |
| Subject: | Re: Postgres Load Profile |
| Date: | 2025-11-04 02:49:01 |
| Message-ID: | CANzqJaBpi-VMu9xOj-qksKoLLoggZNTi13ZUu5eea8M5Rc6eow@mail.gmail.com |
| Views: | Whole Thread | Raw Message | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-admin |
On Mon, Nov 3, 2025 at 5:56 PM Sam Stearns <sam(dot)stearns(at)dat(dot)com> wrote:
> Howdy,
>
> Does Postgres have any tables you can query to find out information such
> as:
>
> - Logical reads
> - Block changes
> - Physical reads
> - Physical writes
> - Read IO requests
> - Write IO requests
> - Read IO (MB)
> - Write IO (MB)
>
>
https://www.postgresql.org/docs/17/monitoring-stats.html
>
> - User calls
> - Parses (SQL)
> - Hard parses (SQL)
> - Executes (SQL)
>
> Probably not unless you want to set log_statement=all and then
parse log_directory/log_filename.
>
> - Transactions per second
>
> Does TPS make any sense beyond when DBMS is running a "strict monoculture"
application like OLTP with *zero* report generation? Because a five hour
SELECT that joins 42 tables in addition to 37 hairy subqueries is just as
much a transaction as is a 5 microsecond SELECT of one customer's records
using a hash index, and an equally fast INSERT of three records.
--
Death to <Redacted>, and butter sauce.
Don't boil me, I'm still alive.
<Redacted> lobster!
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Laurenz Albe | 2025-11-04 05:25:35 | Re: Postgres Load Profile |
| Previous Message | Sam Stearns | 2025-11-03 22:56:27 | Postgres Load Profile |