From: | Hannes Erven <hannes(at)erven(dot)at> |
---|---|
To: | pgsql-general(at)lists(dot)postgresql(dot)org |
Subject: | Big variance in execution times of simple queries |
Date: | 2022-01-24 22:07:02 |
Message-ID: | e2b65bc7-64a3-64fe-dd49-ac316d74c0a0@erven.at |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-general |
Hi community,
I'm looking at a "SELECT * FROM pg_stat_statements" output and am
puzzled by the huge differences between min/max_exec_time even for
simple queries.
The most extreme example is probably the statement used by the
application's connection health check:
SELECT 1
min=0.001, mean=0.00386, max=36.812
Other statements with huge variances include:
SET application_name=$1
min=0.002, mean=0.005, max=9.177
SELECT * FROM table WHERE id=$1 (where ID is the primary key column;
table has 0.5M rows and is frequently vacuum analyzed)
min=0.010, mean=0.260, max=12338.665
According to the system's monitoring, there is no pressure on any
resource (cpu/mem/io). It's 13.5-2pgdg20.04+1 on Ubuntu 20.4; the VM has
12 cpus/16GB memory, ceph-based SSD storage (latency ~1.5ms), and runs
on max_connections=100 with usually 25-40 processes being connected.
Is this to be expected?
Is there something I can watch out or monitor for?
Thank you for any insights...
Best regards
-hannes
From | Date | Subject | |
---|---|---|---|
Next Message | Tom Lane | 2022-01-24 22:57:00 | Re: Big variance in execution times of simple queries |
Previous Message | David G. Johnston | 2022-01-24 17:59:45 | Re: Cannot find hstore operator |