Disabling monitoring after it has been enabled...

[Sean Alderman]

Al,
  Do you have a process for disabling monitoring of a service(s) on a host after they have been monitored?  I've attempted to do this and this is how it seems to go...

1. set service_monitor = false
2. run puppet agent on target host
3. run puppet agent on nagios host
    a. this causes the auto.d/services/host-service.cfg to be chown/chgrp/chmod root:root:0755
    b. this does not cause a nagios service reload
4. manual restart of nagios service does nothing, service monitors are still active.
5. manual removal of auto.d/services/host-*.cfg and execute puppet run redeploys service monitor files and reloads nagios.

In one particular case, including the puppet-tftpd module using xinetd mode leaves behind the faulty xinetd process and tcp monitor - which I can't figure out how there's an xinetd tcp monitor in the first place, except that the tftpd module makes it deploy.

Is there a safer, more automagical way to disable a previously enabled monitor?

Thanks

[Sean Alderman]

Also it would seem like it is difficult to completely remove a host from Nagios even after a puppet node deactivate host has been run on the puppet master.  Manually removing the auto.d/services/host-*.cfg's and auto.d/hosts/host.cfg files is futile after the next puppet run on the nagios server.

Is there a better way for that too?

[Alessandro Franceschi]

Ideally the monitoring stuff should remove checks when the relevant class is o more monitored.
Probably there's a bug, which does not manage correctly your case (disabling monitoring on an existing monitored service) and I have an idea on where it is.
Please open an issue on github on the relevant modules.
I fear it's a common bug, cloned multiple times ...