Downtime, triggered downtime for children and alert sounds

[jpsandiego42]

Hello Folks,
New to the forums, but been using Nagios for a while and dealing with something annoying that I thought I'd ask how to really handle:

I am using parent-child relationships for our hypervisor to guest VMs and this is helping to only send one email notification when a hypervisor goes down, so I think things are setup correctly, but I can't seem to figure out what to do in the UI when we are working to recover a server. When I see a hypervisor go down I setup a downtime for that host hypervisor and set trigger downtime for the child nodes and that stops additional email notifications from being sent, but in the UI, I still get the 'critical' sound playing repeatedly.

What am I missing in order to get the UI to shutup while we're working on the issue so I don't need to just ignore all sounds?

Thanks!

[abrist]

What am I missing in order to get the UI to shutup while we're working on the issue so I don't need to just ignore all sounds?

Thanks!

What are you using that makes sound on alerts? Neither XI nor core does so . . . .

[jpsandiego42]

This installation is Nagios Core 4.0.6. I got some wav files and enabled sounds in cgi.cfg:

Code: Select all

host_unreachable_sound=hostdown.wav
host_down_sound=hostdown.wav
service_critical_sound=critical.wav
service_warning_sound=warning.wav
service_unknown_sound=warning.wav
#normal_sound=noproblem.wav

It has been nice to have the sounds active in the background, but it's really annoying when there's a number of services that go down together...

[abrist]

Ummm,
I think this is a have your cake and eat it too situation. It you want sounds for alerts expect many sounds when you have many alerts.
You are of course welcome to open a feature request on github for more control over the sounds:
https://github.com/NagiosEnterprises/nagioscore

EDIT: The cgi sound options are ancient. I guarantee changes have been made without consideration to that bit of code.

[jpsandiego42]

The issue is not to do with the sounds - that's more the annoying outcome of how child services are handled in a triggered downtime on their parent.

Can it really be true that the parent<->child relationship actions cannot handle that situation? I would expect that if I setup a triggered downtime for a parent host, all warnings for parent services, child hosts and child services would be be automatically acknowledged so they are not considered outstanding (and then cause the sound to be played).

It does seem to work correctly when the parent node goes down completely as it's a blocking outage. The problem comes up when I have a flapping parent and I set a downtime for that parent - the children still continue to complain... I get the feeling I'm just not following the concept of operations for this feature..

[sreinhardt]

Putting a parent into downtime will not effect its children, and in neither case will alerts generated before a downtime be automatically acknowledged. We need to make sure that if nagios saw an issue and it was not found to be corrected with a host\service check during or after the downtime, that you are still informed of this. As for parents going into downtime and not bringing parents as well, there are many reasons for this. Largely due to the idea that if a parent is going into downtime, it is entirely possible that you may not actually want the children in downtime as they have other services that should still be checked and alerted for. Nagios tries very hard to not assume what you as an administrator actually want it to do, and instead expect that you give it explicit instructions. That being said, I think it would be an excellent feature request to have the ability to include child hosts and services in downtime when putting a parent into downtime, this would be very similar to allowing a host to go into downtime and bring the services with it, which already exists.