Number of host/service checks Nagios can handle?

[TBT]

Does anyone want to share information on:

- How many hosts/services you're monitoring?
- Single or multiple Nagio servers?
- Hardware specs of these servers?
- Virtual or physical server?
- Using mod-gearman to do the work?
- Overall how much Nagios can handle?

[lmiltchev]

These are our general guidelines on the hardware requirements needed to run Nagios XI:

http://assets.nagios.com/downloads/nagi ... ements.pdf

[mikew]

This statistics are based on servers I have helped set up in real world situations. So they reflect realistic numbers if you avoid some of the mistakes that impact performance.

There are several issues to consider that do make a great deal of impact on performance:
* number of interface checks
The interface or bandwidth checks are typically the most expensive, take up the most resources. So, if you are primarily checking router and switch ports your numbers will be lower.
* Turn off checks for unused switch ports or make them administratively down and modify ifoperstaus plugin to be OK state for adminDown
* Errors and Warnings
If your organization allows errors and warnings to build and you do not fix them they will take your resources away and your numbers will be lower.
* frequency of checks
Of course if you check every 10 seconds it will take more resources than if you check every 10 minutes.
* Compiled vs. Scripts
Plugins that are compiled use way less resources than perl, bash, etc. scripts. If you can use the default plugins to do the job.
* Passive checks can save you a fortune in resources.


Nagios Physical Hardware
Your bottleneck will be CPU...especially if you do the performance enahancements for IO.

under 2000 checks -- 4 Core, 6 GB of RAM
2000-5000 checks -- performance enhancements, 8 Cores, 8 GB of RAM, DNX or Mod-Gearman
5000-10,000 checks -- performance enahancements, 16 Cores, 16 GB of RAM, DNX or Mod-Gearman
10,000+ -- offload MySQL, performance enhancements, 16 Cores, 16 GB of RAM, DNX or Mod-Gearman


Your results will be somewhat less using virtual machines, but with a virtual machine you can take a snapshot of the container for a backup which is a huge advantage.

The bottom line, a well thought out and managed Nagios server can perform 50% better than a disorganized poorly managed server. Think through what you are doing, clean up your errors and warnings. No two servers will be the same.