BAM instance monitoring

I would like to use BAM to provide violations for process instances where a steps processing time has exceeded a value.

I have set this up with a filter and also used the “event” type for the rule. I can also see related instances when the rule is fired.

However, I’m not sure/convinced that this is acting on an instance by instance basis. BAM tends to want to do everything on aggregations and I know that the “event” type doesn’t wait for that but uses a reading value. However, what is not explained is where the reading value comes from.

The reason this is important is that in some scenarios I would like the rule to be sticky but I want it to be cleared for a specific process instance and not all contributing instances.

The main scenario is that I need to configure alerts that differ by customer, and by specific data that is in the process instance (not aggregated). For instance dates that are specific to a given process invocation.

Any help would be appreciated