Service Dependencies For Windows Services

[cjutting]

Hey Guys First post here and hopefully someone can point me in the right directions.

We have quite a few Windows Based Servers that we monitor countless services on them. We aren't always as good as we should be about putting services in maintenance mode therefor we wind up with tons of text messages / emails to the techs.

I'd like to cut back the number of text messages to the phones using service dependencies very similar to how we are doing it with host dependencies. I have host groups setup for all of my windows based services and all of these servers have and in our case require the NSClient++ to be alive so my thought would be to go down that street, but every rendition I come up with does not act like I want it to..

Does anybody have a good example they could post?

[slansing]

We're more than happy to take a look at your configurations for dependencies and point out where they might be a bit wrong, what is the issue that is currently happening, are the dependent services just not behaving as they should as described here?:

http://nagios.sourceforge.net/docs/3_0/ ... ncies.html

[cjutting]

I'm wondering if I'm mis-understanding the service dependencies.. What I want is if the NSClient++ service is down do not notify me of the dozens of other services that are there.

This is what I have now:

define servicedependency{
hostgroup_name windows-devices
service_description NSClient++
dependent_service_description *
notification_failure_criteria u
}


Notify me for Critical Events.

[slansing]

It should be default that if the host goes down, the services under it are suppressed from sending notifications. NSClient++ is not doing the alerting, Nagios is.

I also do not believe a wildcard will work in dependencies, you would need to name a dependent service, or a group of them.

[cjutting]

If the host goes down Nagios suppresses Services under that host.. That part works correctly, but in some instances we still have some old old old servers that nagios can ping them, but can't read the services as fast as it should so we get notified. Occasionally the NSClient++ service dies and then sets off the rest of the services too.

[lmiltchev]

Have you tried increasing the plugin's timeout value ("-t" flag)?

[cjutting]

I have not. I'm going to go on a limb and say this isn't going to work like i'm expecting.

[slansing]

Well, you could set it up so your services are dependent on a ping check service, which would essentially do the same thing as a host's ping check, just at perhaps a faster rate, or ran in between host checks. Then, when your services notice that this one is in a critical state, they will behave as you are hoping they will. You could also, of course, decrease the amount of time between host ping's of that host, so it could catch that the host is unreachable potentially before the services start failing.