Re: [PERFORM] Strange performance degradation

From: Matthew Wakeling <matthew(at)flymine(dot)org>
To: Lorenzo Allegrucci <lorenzo(dot)allegrucci(at)forinicom(dot)it>
Cc: pgsql-performance(at)postgresql(dot)org, pgsql-general(at)postgresql(dot)org
Subject: Re: [PERFORM] Strange performance degradation
Date: 2009-11-20 15:08:00
Message-ID: alpine.DEB.2.00.0911201500570.684@aragorn.flymine.org
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-general pgsql-performance

On Fri, 20 Nov 2009, Lorenzo Allegrucci wrote:
> performance is degrading...

> In normal conditions the postgres process uses about 3% of cpu time
> but when is in "degraded" conditions it can use up to 25% of cpu time.

You don't really give enough information to determine what is going on
here. This could be one of two situations:

1. You have a constant incoming stream of short-lived requests at a
constant rate, and Postgres is taking eight times as much CPU to service
it as normal. You're looking at CPU usage in aggregate over long periods
of time. In this case, we should look at long running transactions and
other slowdown possibilities.

2. You are running a complex query, and you look at top and see that
Postgres uses eight times as much CPU as when it has been freshly started.
In this case, the "performance degradation" could actually be that the
data is more in cache, and postgres is able to process it eight times
*faster*. Restarting Postgres kills the cache and puts you back at square
one.

Which of these is it?

Matthew

--
Reality is that which, when you stop believing in it, doesn't go away.
-- Philip K. Dick

In response to

Browse pgsql-general by date

  From Date Subject
Next Message Doug Sampson 2009-11-20 15:51:58 PLPerl - Undefined symbol "MemoryContextSwitchTo"
Previous Message Tom Lane 2009-11-20 15:05:55 Re: how to use a cursor for update?

Browse pgsql-performance by date

  From Date Subject
Next Message Kevin Grittner 2009-11-20 15:24:17 Re: Postgres query completion status?
Previous Message Jonathan Foy 2009-11-20 15:01:35 Re: View based upon function won't use index on joins