With the eValid solution server loading is divided up into two parts: Creation of the script(s) to apply to the server, and applying them to observe performance changes?
If you already
have a script, then after that point the methology remains constant: Create a LoadTest script (an *.evl script), run it on one or more machine/user instances, and collect the performance data.
But seem to be asking if this approach has any problem with SalesForce? As far as we know, no, once your SalesForce test script is working well then the rest of the work is pretty routine.
Our experience is that the hard part is to get a good script going in the first case. Here is a YouTube Video of an eValid SalesForce Demo
that shows a complex SalesForce example. This will give you an idea of what's involved in getting a good SalesForce script.