meta-trigger if trigger fires too much

[troffasky]

Is there a way to create a trigger about the state of a trigger? I would like a trigger if a trigger fires so many times within a window. This is sounding a bit like triggering on the content of the "top 100 triggers" report.
The social problem which I am seeking a technical solution to is that colleagues are lazy and will close a ticket as soon as a trigger clears, without reviewing trends.
I am aware of recovery expressions, but we like to use Zabbix to answer "is this up or down?" and if we are waiting some period for the recovery expression to become true then the trigger is "lying".

[kyus]

Have you tried creating a trigger with trend functions?

You could also try creating a trigger with a longer evaluation period. Let's say you have an item that every minute returns 0 when a service is down and 1 when service is up.
If you already have a trigger like:
last(//service.status)=0

To see if the service is down too ofter you may test this:
sum(/host/agent.ping,1h)<30

This would mean that the service was down for half the time in the last hour.

Maybe that'll work for you!

[troffasky]

The problem with using the average over a window to fire is that it does not fire quickly enough - this effectively masks the issue until you're far enough in to the window. That's not what I am trying to achieve. At least with a distinct recovery expression, you can fire and recover fast if it really is a "blip". And if multiple blips occur in the recovery expression window, then it will not clear until all blips have passed.
No, I have not looked at trend functions. I will.