Nagios Support Forum

As i described multiple times before, we have a distributed monitoring setup which uses both the core and xi versions to get all data.

At this moment i am having more then 200 checks per host and about 10 hosts per customer. Because of this amount the services overview gets really big in the xi version. To solve this problem i was thinking of removing the services per customer and go for a global services which would mean only 200 services instead of 200 x hosts x customers.

What i am wondering now is which data gets send to the central itself. In particular the info regarding the hostgroups/templates. Are these being send as well or is it purely the host,service + performance data?

Also when i do this how can i best solve the customer access to the hosts/services. At this moment i have it set so the customer has been added to the host and service template. My idea is to keep the host template and make 1 generic service template but i am wondering if the customer does not get too much access that way. Any ideas on this perhaps?

I'm assuming you're using the Unconfigured objects page to add the hosts and services that are getting pushed to the central server? If so, I believe it's just a basic host and service definition that gets added to the central server, with the xiwizard_passive_host and xiwizard_passive_service templates applied to them. After the passive configs are initially added, the only thing sent to the central is plugin output and performance data.

I have tried the unconfigured objects but the problem there is that it adds the service under the hostname which means in our case that we get (as example) 10 customers with 10 servers each and each server has 200 checks. This means the services screen is hell to browse through. If i can add the 200 services only once and add each server from each customer to it, it will be a bit more work but if my co-worker decides the names are incorrect (which happens quite a lot), then i only need to edit 1 name instead of searching for all x amount.

The only question remains how i can solve the user problem which i described in the first post as well.

If i can add the 200 services only once and add each server from each customer to it, it will be a bit more work but if my co-worker decides the names are incorrect (which happens quite a lot), then i only need to edit 1 name instead of searching for all x amount.

Would it fit for this situation to have your services assigned as a service->hostgroup relationship. That way any new hosts that get added to the group automatically get the services added?

The only question remains how i can solve the user problem which i described in the first post as well.

It's a little bit tricky to answer this question precisely without knowing exactly what you want your customers to be able to see and do, but it might be good to know that if you define your contacts at the service level, instead of the host level, they will only be able to see and access the service information, and not the hosts. You can also control if users can submit commands for hosts or services as well. This can be configured in their contact definition in the CCM.

Hopefully I answered your question ok

mguthrie wrote:

If i can add the 200 services only once and add each server from each customer to it, it will be a bit more work but if my co-worker decides the names are incorrect (which happens quite a lot), then i only need to edit 1 name instead of searching for all x amount.

Would it fit for this situation to have your services assigned as a service->hostgroup relationship. That way any new hosts that get added to the group automatically get the services added?

This was my first idea but my co-worker does not want more hostgroups then we currently have. I could see if i can incorperate them in the current hostgroups but then i will have a issue with the 2003/2008 versions of each command.

The only question remains how i can solve the user problem which i described in the first post as well.

It's a little bit tricky to answer this question precisely without knowing exactly what you want your customers to be able to see and do, but it might be good to know that if you define your contacts at the service level, instead of the host level, they will only be able to see and access the service information, and not the hosts. You can also control if users can submit commands for hosts or services as well. This can be configured in their contact definition in the CCM.

Hopefully I answered your question ok

We want the customers to see there own hosts and services but nothing else. At the moment i have a service template for each customer where i add there contact. From there they can see there hosts and also services. The only downside is that i have to check this with every update that is being released of course. Also i am not sure if this is the best way to go.

Keep the contact information in a separate template in a multi-tenancy situation is probably the best way to go. As far as the updates go, we don't do anything to touch anyone's existing monitoring configuration, so nothing should be changing for contacts from one release to the next.

Ok, thanks for the help.