nagios for large installations

[cresg82]

hi,
i have a question about nagios performance and capabilities. I need to monitor 2000 hosts with ad average of 30 services per host=60.000 services. It is possibile to do this with nagios core?
how many hosts and services can manage with nagios core? What are the hardware requirements?
thanks

[agriffin]

You'll need a physical server with a fast CPU and fast hard disks. You'll also probably need to distribute the checks using something like mod_gearman or dnx. Unfortunately it's hard to get very specific about requirements since everybody's use case for nagios can vary quite a lot.

[cresg82]

ok...suppose i have a virtual server on an esx host with 2 cpu, 4 gb ram and 80 gb ok disk space... how can i calculate how many hosts/services can i manage of an istance of nagios installed on my VM

[agriffin]

There really isn't a good answer I can give you. It depends not only on your hardware, but also the specific checks you will be using.

[jsmurphy]

hi,
i have a question about nagios performance and capabilities. I need to monitor 2000 hosts with ad average of 30 services per host=60.000 services. It is possibile to do this with nagios core?
how many hosts and services can manage with nagios core? What are the hardware requirements?
thanks

agriffins advice is sound, Nagios can be notoriously difficult to calculate the scalability of due to the fact it relies heavily on third-party plugins... and that is where the majority of your performance hit is going to be. Some plugins are miniscule in their utilization and some (like the VMWare API) will rape and pillage your server of every last resource it can get its hands on.

Assuming however a basic setup of the 'big three' SNMP, NRPE and NSClient, you would probably be pushing a single server to near its limits with that number of services but it could probably be done with adequate optimization. The smart money however would probably be running two boxes with DNX for the sake of future scalability.