Selective notification for services in hostgroup

[rajasegar]

Nagios XI 2012R2.9
RHEL 6.5 x64
Manual Install

I have created a bunch of services common to Windows and assigned a hostgroup that has all the windows servers in it.
It works great as I dont have to create new services for each new server.

However then comes the following scenarios
1) Some windows servers need different thresholds for say Drive C, however my service has a common threshold
2) For some servers, Nagios should notify Contact A for Sql Server services issue and Contact B for capacity issues.
If I put a contact in these common service, they will get notified for all servers since it is common.

What is the best way to go to achieve this?
Last thing I want is to duplicate services for each host and have a nightmare if there is a need to do global changes.

Thanks in advance.

[lmiltchev]

You have one service, that is assigned to a hostgroup. It is not possible, as far as I know, to modify it for one member of the hostgroup, but not for another. It's still the same service. Same with notifications. If you don't have way too many checks that need to "differ", you can create a separate checks for these services. Unfortunately, this seems to be the only way.

[tmcdonald]

There might be a middle ground with templates, but I think you would still have a pretty complicated setup with multiple groups.

[rajasegar]

Thanks. I will avoid all the issues and duplicate the services for each host.
Will there be any performance issues related to this?
We have about 1000 servers and each about 10 - 15 services to check.

[Box293]

I've faced this issue in the past and I resolved it with multiple hostgroups.

With the disk space checks stuff, I orgranised the hosts into a couple of host groups:

• Servers with disks that have less than 200 GB
  Servers with disks that have less than 600 GB
  Servers with disks that have less than 1 TB

Then you create a a service for each of these host groups with the relative thresholds. Sure you end up with three services instead of one but it's better than one service per server.

This is one way to approach the problem.

[rajasegar]

I've faced this issue in the past and I resolved it with multiple hostgroups.

With the disk space checks stuff, I orgranised the hosts into a couple of host groups:

• Servers with disks that have less than 200 GB
  Servers with disks that have less than 600 GB
  Servers with disks that have less than 1 TB

Then you create a a service for each of these host groups with the relative thresholds. Sure you end up with three services instead of one but it's better than one service per server.

This is one way to approach the problem.

I had similar thought to use common services for all expect for those that need different thresholds.
Then there is the issue of how to handle various different notifications. I have to create different services for this also.
In the end there will be a mixture of configurations here and there and will be very hard to keep track and manage.
End of the day this must be easily supported and maintained especially with very large setups.

Thanks for sharing your experience on this.

[tmcdonald]

I think @Box293 has the best solution.

The unfortunate thing is that with increased granularity comes increased complexity. By default, Nagios XI tends to create one separate service entry per service, as opposed to just one and applying that to many hosts. There is certainly a middle ground between the 1:1 and 1:many relationship, but ironically you end up creating something more complex by trying to simplify it. Once you settle on a solution though, and stick with it, things tend to get easier going forward. Just a lot of up-front effort.