Good point.
Presumably one could monitor the field and avoid taking data during the satellite pass. Could be an interesting exercise writing software for motion detection in the autoguider camera.
How do you deal with satellite trails now?
Depending on whether I am doing photometry or imaging (which includes astrometry for present purposes) I either discard the sub before average-stacking or rely on median-stacking to discard the satellite for me. In either case, such concerns are of interest only when the satellite passes inconveniently close to the objects of interest, otherwise it is just ignored.
There again, I take as many 30 to 60 second subs as are needed to reach a satisfactory SNR (which depends on sky brightness, for instance) and then move on to another target. A cheap and simple way of optimizing the productivity of telescope time. Vary rarely does a single exposure exceed a couple of minutes.