Load Balancing WebGUI
Posted: Fri Feb 17, 2017 5:27 am
Hi all.
We are trying to speed up our Nagios WebGUI and have done just about everything we can to achieve this.
We have an environment where we have one main Nagios server and all checks are done by 6 mod_gearman nodes.
So our main Nagios server isn't really doing any checks, its just accepting the check results and processing them.
We have moved our main Nagios server to an all flash VM environment for faster disk access.
The number of CPUs has been doubled from 2 to 4
There is 24G RAM and 50G swap
But the WebGUI isn't a quick as some people would like.
We are monitoring around 2500 hosts and 15000 services, all of which are done on mod_gearman nodes
We are running Nagios Core 4.2.4.
The main load seems to be when people are access status.cgi for check info and submitting commands from the GUI.
Running top and watching the individual CPUs I can see most are running at about 30%, then they will jump to 90% and the process on the stop is status.cgi.
We have a lot of people accessing the WebGUI which is why this is happening.
Just wondered if anyone has found a way of load balancing Nagios across more than one server and still have identical information in the output of the WebGUI and are able to run commands from the GUI.
Or if anyone had any suggestions?
Thanks
Tony
We are trying to speed up our Nagios WebGUI and have done just about everything we can to achieve this.
We have an environment where we have one main Nagios server and all checks are done by 6 mod_gearman nodes.
So our main Nagios server isn't really doing any checks, its just accepting the check results and processing them.
We have moved our main Nagios server to an all flash VM environment for faster disk access.
The number of CPUs has been doubled from 2 to 4
There is 24G RAM and 50G swap
But the WebGUI isn't a quick as some people would like.
We are monitoring around 2500 hosts and 15000 services, all of which are done on mod_gearman nodes
We are running Nagios Core 4.2.4.
The main load seems to be when people are access status.cgi for check info and submitting commands from the GUI.
Running top and watching the individual CPUs I can see most are running at about 30%, then they will jump to 90% and the process on the stop is status.cgi.
We have a lot of people accessing the WebGUI which is why this is happening.
Just wondered if anyone has found a way of load balancing Nagios across more than one server and still have identical information in the output of the WebGUI and are able to run commands from the GUI.
Or if anyone had any suggestions?
Thanks
Tony