I agree that a distress scale might be useful. This could be completed almost contemporaneously if available on a smart phone or tablet for subjects. If you want to measure stress immediately I would suggest a biological monitoring mechanism to record heart rate, muscle tension or skin conductance. You could also get subjects to provide subject qualitative reports on stress indicators afterwards by voice recording and then do content analysis on this. Each method would give a different type of data and the choice I guess depends on what outputs you want from the study. A mixed methodology is likely to be most interesting scientifically and practically. I would suggest minimising the lag time between the event and measurement of stress whatever method is used and also involving subjects in the design and testing of the method through cognitive testing of any questions, piloting etc, and acceptability of any devices used.
See our recent related work on quantifing feeling and its link to mobile use.http://ieeeaccess.ieee.org/featured-articles/notimind-utilizing-responses-smart-phone-notifications-affective-sensors/