Timeout on TICK alerts to force them to clear (not just go into recovered state).

[docmerlin]

Subcomponent
TICKscript

Feature Request Summary
Someone is currently doing network monitoring of multiple sources/metrics - mostly via telegraf doing SNMP, ping etc.
If a device goes down, they get an alert.

The use-case they're trying to satisfy is:
If a device is decommissioned, and someone forgot to remove it from monitoring (in Telegraf config), the alert will fire but will never recover. They want these alerts to time out after (say) 2 days.

They would like barrierNode to be able to delete events, such that if they did something like:

|> alert()
.blah
|> barrier()
.idle(2d)
.delete(TRUE)
the event created in alert() would get deleted.