We have set up a monitor which notifies us of potential performance issues for each tenant in our system.
However, there are days where activity might be very low on some specific tenant, and the request count may not be worthy enough of triggering an alert.
For example, the monitor should only trigger on BOTH of the following:
- if requests' duration p50 over 24 hours exceed 100ms, AND
- if request count is greater than 100 in 24 hours.
How can I achieve the request count condition? There seems to be no way to do this out of the box, but is there a creative way to go about it?
