Thank you for your feedback, but in this case these won't help, but let me clarify the situation, since most of the time these suggestions would be absolutely spot on. 

In my case at the moment I have an all purpose interactive cluster for my developers to use, which means it's usage is very ad hoc and it's used by a growing number of users (half a dozen now, but we intend to at least double that in the short term and to keep adding) .

Now with the addition of the serverless clusters that we can use to run our notebooks I wanted to be able to more accurately predict how much switching to it over the all purpose  would cost, but comparing just a straight comparison of how much time I have the cluster up and how much it would cost a serverless for the same amount of time is not very accurate and gives a skewed picture, so I was trying to find a why to have a more accurate idea of how much serveless usage we would have.