Parent / Child relationships in concern to notifications

[sburckhard]

Nagios XI 2009 R 1.3G
CentOS 5.5

I am having a problem with parent / child relationships.
When I specify the parent, it correctly displays it in the network status map, but alerts do not seem to be aware of the parent/child relationship.

Example:
Switch 1 has 6 hosts behind it
when Switch 1 goes down I get emails for all 7 devices, rather than Just switch 1,

Also last night I had a site go down
Router, switches, servers, and services monitored all generated alerts, so I had over 100 emails...

I am under the impression that if a "parent" goes down, it suppresses the email notifications for all of its children, or am I wrong in that thought?

[mcleodc]

Hi,

Well it depends on what notification settings you have enabled. The parent child relationships you have set up should allow Nagios to decide if a host is down or just unreachable because its parent is down. With the parents set in each host when a site goes down you should expect to receive a host down email about the router ect. and host unreachable alerts about the rest of the hosts. Are the alerts you receive host unreachable or host down alerts?

You can turn on/off unreachable notification on a per contact basis or change it in the contact template.


Thanks
Chris

[sburckhard]

Basically when a host goes down, I get service alerts for every service I monitor on that server.

Which seems odd, If the host is down, it should only send a host email, the services are a given, if the host is down of course all the services are down...

I want to be alerted if a service fails, but I do not need 12 service alerts when the host has gone down, I just need that host notification.
I am wondering because it went into a "flapping" state if that triggered the storm of emails.

[sburckhard]

I think I figured out what you mean now...

[mcleodc]

Hi,

I have also setup service dependencies for each device. So for a windows server I would have all services monitored being dependant on the ping and nsclient service, this stops all the extra service alerts.

this is a copy of the service depenencies that the NagiosXI gui created:


define servicedependency {
dependent_host_name DRUAS1
dependent_service_description CPU Usage,Drive C: Disk Usage,Logon Errors,Memory Usage,Page File Usage,Server Work Queues,Uptime
host_name DRUAS1
service_description NSClient,Ping
inherits_parent 1
execution_failure_criteria w,u,c
notification_failure_criteria w,u,c
dependency_period 24x7

Thanks
Chris

[rdedon]

Hi,

I have also setup service dependencies for each device. So for a windows server I would have all services monitored being dependant on the ping and nsclient service, this stops all the extra service alerts.

this is a copy of the service depenencies that the NagiosXI gui created:


define servicedependency {
dependent_host_name DRUAS1
dependent_service_description CPU Usage,Drive C: Disk Usage,Logon Errors,Memory Usage,Page File Usage,Server Work Queues,Uptime
host_name DRUAS1
service_description NSClient,Ping
inherits_parent 1
execution_failure_criteria w,u,c
notification_failure_criteria w,u,c
dependency_period 24x7

Thanks
Chris

Thanks Chris for posting these examples, we appreciate the help! sburckhard if you are still having issues or need some additional guidance we will be more than glad to help!