Hi,
I'm looking at a large NAGIOS XI implementation and need to get some idea of the resource requirements.
I have the option of running as a virtual guest on VMWare ESX or as a physical server.
Obviously a physical server costs more than a virtual but does NAGIOS need a physical or can it run in a single/multiple core guest and what memory/network does it need?
Do you have a sizing document for NAGIOS XI?
Thanks
XI capacity requirements
-
tonyyarusso
- Posts: 1128
- Joined: Wed Mar 03, 2010 12:38 pm
- Location: St. Paul, MN, USA
- Contact:
Re: XI capacity requirements
While the actual requirements vary greatly as mmestnik referred to, I can confirm that it will work in a virtual machine (we actually do much of our internal testing in VMs, and officially distribute a VMware image of XI). The image we publish is an appliance based on CentOS, so the absolute minimal requirements will be similar to that, and scale depending on your needs from there.
The CentOS documentation suggests a minimum of 256MiB RAM, one core or virtual core, and 6GB of disk space per virtual machine.
That said, you can get by with less if you really want to. If you modify the VM image you can get it down to 2.5GB of disk space, and it will likely run okay on 128MiB RAM if you're patient with the web interface. The equivalent of a PIII processor or better or so would be a good idea, again because of the web interface. The actual monitoring and alerts functions don't take very much resource-wise on the server, but configuring and displaying them graphically will need a bit more oomph.
Since you say you're looking to do a large deployment, you won't want to be at those minimums. How it scales will depend on whether your checks are taxing on the server or client side, and how many there are reporting back. A really good idea is to do a partial deployment and graph your resource usage as you add hosts/services and try to project that for your infrastructure (which will also help with your capacity planning in the future).
The CentOS documentation suggests a minimum of 256MiB RAM, one core or virtual core, and 6GB of disk space per virtual machine.
That said, you can get by with less if you really want to. If you modify the VM image you can get it down to 2.5GB of disk space, and it will likely run okay on 128MiB RAM if you're patient with the web interface. The equivalent of a PIII processor or better or so would be a good idea, again because of the web interface. The actual monitoring and alerts functions don't take very much resource-wise on the server, but configuring and displaying them graphically will need a bit more oomph.
Since you say you're looking to do a large deployment, you won't want to be at those minimums. How it scales will depend on whether your checks are taxing on the server or client side, and how many there are reporting back. A really good idea is to do a partial deployment and graph your resource usage as you add hosts/services and try to project that for your infrastructure (which will also help with your capacity planning in the future).