Nagios advanced Configuration - changing contact_groups after defined by hostgroups - nagios

Question

is it possible to set an default contact_group in hostsgroup and modify (redefine) this contact_group later in the service. Or by another way.
Background: We use nagios to monitor near 100 Hosts and near 1000 services. For easy configuration we defined hostgroups with default checks and default contactgroups. But there are more and more services on special server that must have other contactgroups.
Config-Example that shows what we want, but it doesn't work... (Last Part - Redefine)
####################
## HOST GROUP
####################
define hostgroup{
hostgroup_name WINDOWS
alias Windows Server
}
####################
## SERVICE
####################
define service{
use generic-service
name RDP-CHECK-TCP-PORT
hostgroup_name WINDOWS
service_description RDP - Check TCP Port
check_command check_rdp
contact_groups Admin_Alarm
}
define service{
use generic-service
name DF-CHECK
hostgroup_name WINDOWS
service_description HD - Disk free
check_command check_df_win
contact_groups Admin_Alarm
}
####################
## HOST
####################
define host{
use generic-urgent-host
host_name server1
hostgroups WINDOWS
alias Server1
address 1.2.3.1
}
define host{
use generic-urgent-host
host_name server2
hostgroups WINDOWS
alias Server2
address 1.2.3.2
}
define host{
use generic-urgent-host
host_name server3
hostgroups WINDOWS
alias Server3
address 1.2.3.3
}
####################
## Redefine Contactgroups
####################
define service{
host server1
service_description RDP - Check TCP Port
contact_groups No_Alarm
}
define service{
host server2
service_description RDP - Check TCP Port
contact_groups Urgent_Admin_SMS_Alarm, Customer_Admin_Alarm
}

Answer 1

contact_groups can only assigned to hosts, not hostgroups. However you could use templates and put contact_groups into that templates. Later you can override the contact_groups in your host definition.
Have a look at the objectinheritance docs.

Answer 2

Since this hasn't been said:
There are tools that let you express that as rules and have simple overrides to the default settings.
An escalation should rather be used for something like "I want this to immediately be visible and written to our xmpp chatroom as a notification. But if it stays for 30 minutes broken, we also wanna email the manager so he knows we have a noteworthy outage".
As for tools, I'll recommend Check_MK, which is around since 2009. Initially, it was primed for this task. These days it does a lot more.
Disclaimer: I contributed to it, since I think it is a good thing.