question

I made this example model. It has a unique, but powerful, approach. record-response-time-log.fsm

At a high level, here's the idea. You make a token for each item (using Process Flow), and then fill that token with the data you need, and then, using the Statistics Collector, you log off that data, one event at a time.

Here's the lower level explanation. First, I put the processors in a group. This lets me listen to all processors in that group at once. In my Process Flow, I have an Event Triggered Source. So any time an item enters any of the processors in the group, that source will spit out a token. The token adds a label for which item it has, and which processor it's dealing with. Then, the token sets labels for the current time.

Next, the token waits for the processor to change state to Processing. Once that happens, the token caches of the current time, and calculates the response time.

All the Statistics Collector does is log off all the important token label values, making the table you requested.

NOTE: there are some breathe activities in this flow. This is because there are many, many state changes on the processor that happen in zero time. The purpose of these breathes is to skip any changes to Processing that occur before the item begins processing. This may not work in a future version. If it doesn't, try adding another breathe, or even waiting for a tiny amount of time, like 1.0e-6 or something. And if that still doesn't work, ask a new question, and we'll help you out.