NSCA or NRPE for Performance

Michael S. Kazmier mkazmier at sofast.net
Mon Jul 28 23:41:09 CEST 2003


Hello list,

 

We have been using Nagios for nearly a year now and love it!  We are looking
at extending our utilization of by integrating NSCA/NRPE.  In the
environment that we hope to utilize the remote service, the remote machine
queries nearly 1000 devices via SNMP every 10 minutes.  Nagios by itself
does a great job of scheduling and runs smoothly.    I have a nagios test
machine setup and there is almost no laod (never above 0.2 thanks to the
SNMPPD!!!), however, I am concerned that once I have 100 or more of these
remote machines running, all monitoring 1000 devices (yes, that is about
100,000 devices) the nagios central server will grind to a halt.  Currently
this is a dual Xeon 2.4 with 2 GB of ram.  I realize we may need to segment
these "central" servers to allow for proper scalability.

 

My question is, however, which is better for performance on the central
server, NSCA or NRPE?  I read in one thread about how NSCA sends all data to
a FIFO pipe and that pipe has a limited buffer so service checks which come
in all at once may get dropped.

 

Please advise!

 

Thanks,

 

Mike

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://www.monitoring-lists.org/archive/users/attachments/20030728/960313a9/attachment.html>


More information about the Users mailing list