Re: Distributed NagiosXI with central Configuration manager
Posted: Tue Jan 21, 2014 5:38 pm
It's somewhat hard to guess.
A server that is not doing any processing (running checks, getting snmp traps, etc) and is only receiving and displaying data is going to be able to handle much more than a server which also has to do checks itself. This is just the nature of computation. The biggest bottleneck is going to be a toss-up between generating reports and generating graphs. Absorbing and storing data is easy, however searching it and formatting it is another thing entirely. Unfortunately I personally don't have any useful numbers because I have never seen a single system handle anywhere near 50,000 x 25 = 1,250,000 results every 5 minutes.
There are always the usual things you can do to speed it up (RAM disk, more hardware, etc.) but at some point it will become an issue.
We can run some theoretical numbers and maybe come up with a ballpark, but it entirely depends on the frequency and type of checks.
A server that is not doing any processing (running checks, getting snmp traps, etc) and is only receiving and displaying data is going to be able to handle much more than a server which also has to do checks itself. This is just the nature of computation. The biggest bottleneck is going to be a toss-up between generating reports and generating graphs. Absorbing and storing data is easy, however searching it and formatting it is another thing entirely. Unfortunately I personally don't have any useful numbers because I have never seen a single system handle anywhere near 50,000 x 25 = 1,250,000 results every 5 minutes.
There are always the usual things you can do to speed it up (RAM disk, more hardware, etc.) but at some point it will become an issue.
We can run some theoretical numbers and maybe come up with a ballpark, but it entirely depends on the frequency and type of checks.