Hello,
I'm using Nagios with NRPE to monitor some linux servers via a "proxy" server running NRPE too.
So if the proxy's NRPE stop working, the monitored servers after won't be able to be checked, and Nagios will consider them as DOWN. I will receive 2 mails (1 for the proxy's NRPE service CRITICAL, and one for the distant monitored server "DOWN").
I want to disable notifications for this distant-server if the service NRPE is CRITICAL on the proxy.
I've tried with hostdependencies and servicesdependencies, but it seems that a service & host dependency is not possible..
Any help ?
Edouard
Disable host notification if a specific service is critical
-
- Skynet Drone
- Posts: 2620
- Joined: Wed Feb 11, 2015 1:56 pm
Re: Disable host notification if a specific service is criti
I can't think of an elegant solution.
An inelegant one would be to create a 3rd host - call it proxy, make it's check_host_alive "check_nrpe -H proxyhost" then make it a parent of host 2. The idea being to turn the NRPE service into a host so you have something to work with.
An inelegant one would be to create a 3rd host - call it proxy, make it's check_host_alive "check_nrpe -H proxyhost" then make it a parent of host 2. The idea being to turn the NRPE service into a host so you have something to work with.
Re: Disable host notification if a specific service is criti
Thanks for your reply,
Another solution would be to put the check NRPE command in the hostcheck command for the proxy -> it will send me only one mail like "proxy server is DOWN" instead of "NRPE service is CRITICAL"
I will think about your idea
If anyone have a better solution, please share
Edouard
Another solution would be to put the check NRPE command in the hostcheck command for the proxy -> it will send me only one mail like "proxy server is DOWN" instead of "NRPE service is CRITICAL"
I will think about your idea
If anyone have a better solution, please share
Edouard
-
- Skynet Drone
- Posts: 2620
- Joined: Wed Feb 11, 2015 1:56 pm
Re: Disable host notification if a specific service is criti
Your solution is actually *the right* solution if you can't create a proper (ICMP) host check for the remote host. Of course it can result in false alerts - but in the event that you can't monitor it by any other means they really wouldn't be any more false than to say that it's up all the time.