Impliment alarm flood protection

[bmfmancini]

say in the case where Cacti has lost access to all devices due to a momentary blip in the network
On bigger systems monitoring thousands of devices this can lead to Cacti sending many false tickets to the ticketing system

It would be good to have a mechanism to detect a large influx of down devices and either flag the admin
Or wait a few more mins to see if everything normalizes before sending alarms either by Email or command execution

An admin should be able to configure based on the sizing of their cacti
if I have 1K devices and 500 devices go down in one shot wait 2 more polling cycles for example

[TheWitness]

@bmfmancini, I've got this started, but hit a small snag.