Has anybody ever done this before? if yes, does this one GUI VUser could reflect the performance under 501 loads?
This is the expected use model of GUI Virtual Users. Be sure and label your transactions in the same places as you are measuring them in your API virtual users and then in your results you can examine the performance of LOGIN (API/PROTOCOL Level) and LOGIN_GUI (THROUGH GUI VIRTUAL USER). GUI-API/PROTOCOL = GUI Weight/time.