perl-modperl mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Rolf Banting <rolf.b...@gmail.com>
Subject Re: Profiling live mod_perl backends
Date Mon, 30 Mar 2009 11:46:09 GMT
On Sun, Mar 29, 2009 at 9:52 PM, Perrin Harkins <pharkins@gmail.com> wrote:

> On Sun, Mar 29, 2009 at 4:44 PM, Cosimo Streppone <cosimo@streppone.it>
> wrote:
> > The main problem is that in the past we experienced some kind of
> > performance problems that only manifested themselves really clearly
> > in production and only at peak traffic hours.
> > Out of peak hours, everything was fine.
>
> That sounds like a problem with a shared resource like the database,
> not something you'll find by profiling the code.  You'd be better off
> either using DBI::Profile or using logging on your database to find
> the problem.
>
> - Perrin


There's a neat idea in "Perl Hacks" where you nominate every 100th (or
whatever) object as a debug object. Perhaps you could do something similar
with Devel::NYTProf - profile only every 'n' th request during peak times.

I assume you have checked "obvious" things like cpu & disk usage stats,
KeepAlive, MaxClients etc

Mime
View raw message