[omd-users] Ways to improve scalability

Angel L. Mateo amateo at um.es
Wed Dec 2 12:20:23 CET 2015


El 02/12/15 a las 11:47, Sven Nierlein escribió:
> On 02.12.2015 10:41, Angel L. Mateo wrote:
>> Hello,
>>
>>      We've been running omd for a time, but now we are having performance issues.
>>
>>      We have a OMD server running in a VMware VM with 8 cores and 16GB and load average is over 10 points. This server is monitoring 850 hosts with more than 3000 service checks. We have 9 different sites (one of them has no hosts neither checks, it is used just as a collector of the other 8) running all in the same server.
>>
>>      In order to improve performance we are studying two different ways:
>>
>> a) Moving each omd site to a different VM.
>> b) Using mod_gearman ([1]) so check commands were distributed between different servers
>>
>>      Does anybody has experience using mod_gearman with omd?    What would you recommend? Any other suggestion?
>>
>> [1] https://labs.consol.de/nagios/mod-gearman/
>
>
> 3k services doesn't sound that much. What kind of performance problems do you have?

	We have very high CPU usage in top. Almost 100% of CPU usage with very 
low (most times even 0) idle and wait. All of the CPU load is due to 
user load.

	I'm suspecting now that the problem could be a lot of snmp checks that 
I don't know its performance.

-- 
Angel L. Mateo Martínez
Sección de Telemática
Área de Tecnologías de la Información
y las Comunicaciones Aplicadas (ATICA)
http://www.um.es/atica
Tfo: 868887590
Fax: 868888337


More information about the omd-users mailing list