Using Mod-gearman on Nagios 2012

[MSPk]

Hi Team,

We have setup a PoC environment for one of our clients and we are currently monitoring 10 hosts and 186 services. My Nagios XI server is built on CentOS 6.4 with 4 core CPU, 8 GB RAM and 200GB hard drive. I have installed mod-gearman server on this Nagios server this morning and after few hrs checked the monitoring engine status and I see the below stats.

Host Check Latency
Min 0.00 sec
Max 0.42 sec
Avg 0.14 sec

Service Check Latency
Min 0.00 sec
Max 0.36 sec
Avg 0.12 sec

I'm not quite sure what the latencies were earlier, however I'm trying to understand why we should have a latency for host check considering that we have 4 core cpu and hardly have 10 hosts to monitor...!! Should I be worried about this?

[sreinhardt]

Absolutely not, you should not be worried with one half second max latency and an average of under one fifth. This is pretty great actually. Essentially nagios is queuing your checks, and only executing so many at one time, if one of the other checks takes longer than expected it may wait a period of time(very miniscule) before executing the next check, this is all you are seeing. For example with my own system, dual core and 4gb ram, on a virtual machine with under 100 service checks, and I have a max service latency of 1 second.

[MSPk]

Thats great....

I happened to check the top CPU consumers using the 'TOP' command on my Nagios XI server. I see a process 'check_oracle_health' coming up every in the list every few secs. Shouldn't nagios handle this process to Mod-Gearman server? So, y do I still see this plugin name in the list of processes using the high CPU?

Also, I have used the below documentation to install my modgearman server(when run with --server option it has installed a worker as well), I have jus followed the instructions in there and haven't made any other changes, Please suggest.

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

[scottwilkerson]

Shouldn't nagios handle this process to Mod-Gearman server?

Only of you are not running a mod_gearman server locally, and only if you have installed the check_oracle_health plugin and SDK on the remote mod_gearman worker

[MSPk]

Are you saying that if we are using a mod-gearman worker on a remote host, the Oracle SDK and check_oracle_health Plugin needs to be installed on that remote server as well?

If yes, what should be the location of the plugin? /usr/local/nagios/libexec??

[slansing]

The plugins on the remote worker should reside in the same location they do on the Nagios XI server so that when called from the gearman server, mod_gearman workers can look there for them.