2.0 CGI speed

[Andreas Ericsson]

Jason Martin wrote:
> I've been doing some performance tests of Nagios with the latest
> CVS code and am having some concerns about large-scale
> installations.  For example, with 20K passive services it takes
> 
>>10 minutes on 700Mhz machine to get a status.cgi back. Is there
> 
> anything that can be done to improve upon this? (aside from
> machine upgrades, I'm going to do that anyway).
> 

Put logfiles and object cache on RAM-disk and copy to real disk when 
powering off, or write an eventbroker for keeping recent status in a 
database (backtracking the archives will take forever no matter how you 
do it, due to the sheer amount of data involved).

Actually, I think the smoothest way to go about doing it would be to 
split Nagios in two (I've said this before...) and have one part be the 
logging end and capable of answering questions from the CGI's. With some 
clever hashing (readily available in gdbm) this would be way faster than 
any database, and the possibilities of extending Nagios's logging 
features would be nearly endless.

Just out of curiosity; How on earth did you manage to find 20,000 
services in need of checking in a single network?


> Thanks,
> -Jason Martin
> 

-- 
Andreas Ericsson                   andreas.ericsson at op5.se
OP5 AB                             www.op5.se
Lead Developer


-------------------------------------------------------
This SF.Net email is sponsored by BEA Weblogic Workshop
FREE Java Enterprise J2EE developer tools!
Get your free copy of BEA WebLogic Workshop 8.1 today.
http://ads.osdn.com/?ad_id=5047&alloc_id=10808&op=click