Hosts without a Service
Posted: Fri Nov 13, 2015 3:59 pm
I am trying to remember the rule here and I hope someone can refresh my memory. It has been years since I built our Nagios implementation. It is Nagios 4 Debian and running the current plugins.
My question surrounds the need for creating a service associated with each (or all) host entries created. If I create a host entry alone, nothing would appear in the web interface for that host. I believe it is necessary to create a wildcard service that is automatically assigned to each host to have the host register in Nagios and have its record appear in the Web interface. To that end a while ago I created a "pingall.cfg" object file that uses a wildcard in the host name field to ping every entry. This has always served us well, but now I have a problem. We have a set of machines in international sites that we monitor across a long distance pipe. The ping times exceed our local settings by a great degree. I need to accommodate the long ping times for these special hosts, but now I have the wildcard getting in the way. I can't take it away, I have more than 750 hosts using it as their primary service. I can't make a list of hosts and add it to the ping command, it is 750 entries long and I don't have one, regardless.
Can you think of any way around this? I was thinking of simply creating a second implementation of Nagios for those international machines. That might be my best bet.
How do you folks get around the need for a service for each host? Do you create one globally as I did, or do you build one in each cfg file explicitly? I'd appreciate knowing.
Thanks
JB
My question surrounds the need for creating a service associated with each (or all) host entries created. If I create a host entry alone, nothing would appear in the web interface for that host. I believe it is necessary to create a wildcard service that is automatically assigned to each host to have the host register in Nagios and have its record appear in the Web interface. To that end a while ago I created a "pingall.cfg" object file that uses a wildcard in the host name field to ping every entry. This has always served us well, but now I have a problem. We have a set of machines in international sites that we monitor across a long distance pipe. The ping times exceed our local settings by a great degree. I need to accommodate the long ping times for these special hosts, but now I have the wildcard getting in the way. I can't take it away, I have more than 750 hosts using it as their primary service. I can't make a list of hosts and add it to the ping command, it is 750 entries long and I don't have one, regardless.
Can you think of any way around this? I was thinking of simply creating a second implementation of Nagios for those international machines. That might be my best bet.
How do you folks get around the need for a service for each host? Do you create one globally as I did, or do you build one in each cfg file explicitly? I'd appreciate knowing.
Thanks
JB