Nagios Support Forum

Hi Team,

What is the optimistic hardware required to monitor 1000 hosts with 12k check . I would need for both physical and virtual hardware details.

Physical :

Hard Disk Space
CPU Cores
RAM

VM:

Hard Disk Space
CPU Cores
RAM

Please fill in your recommendation for the above physical and VM.

For 12k checks - and I am making certain assumptions here (5-minute check intervals, mostly compiled plugins, etc) - I would do at least 200G of disk, 8 cores and 8GB of RAM. There are a ton of variables at play when it comes to provisioning a server, so the best I can give is an estimate. At any rate, at around the 10k check mark I usually recommend implementing a RAM disk at the least to reduce IO, and offloading the database:

https://assets.nagios.com/downloads/nag ... giosXI.pdf
https://assets.nagios.com/downloads/nag ... Server.pdf

You can also look into implementing mod_gearman to offload checks to other servers:

https://assets.nagios.com/downloads/nag ... ios_XI.pdf

Physical vs virtual should not make much difference in regards to the specs, with the exception that too many vCPUs on a VM can actually be detrimental: http://www.gabesvirtualworld.com/how-to ... rformance/

Hi ,

If I need to offload the database what should be the hardware specification of the database server ?

Which ports should be open between the Nagios server and database server for connectivity ?

Under which mount point does the database information being saved by default ?

I would say the database certainly needs less resources than the XI system itself, but I do not have any benchmarks for this specifically. Focus on CPU and fast disks for the best performance, and I would not go below 4GB of RAM. 100GB of the 200GB disk I estimated for the XI server itself probably should go to the database.

By default, MySQL listens on port 3306 but this might not be the case in your environment. Whatever port the offloaded database listens on, that is what must be opened for the XI machine to communicate.

Also by default, our XI MySQL databases install directly to /, but again this might not be the case in your environment.

We are requesting for below hardware for hosting Nagios XI.

8 CPU, 16 GB RAM, Storage of 300 GB for monitoring 500 servers with around 5000 service checks.

In addition to above detail we need to also provide IOPS details .

As per your experience can you please suggest what would be the IOPS (input/output per second) which I need to request for ?

Server provisioning is somewhat out of scope for the support forums, and is moving into consulting. I don't have an estimate for IOPS - that number, as with most metrics, will likely vary based on what sort of checks you have, how often they run, etc. The iostat command will show you the tps (transfers per second, same as IOPS) value which is what you would need to measure, but I don't have information on what sort of numbers other users would be seeing.