I am pretty sure this is not currently present in SGP but if something like this does exist, please point me to it. If not, consider this a feature request.
Presently my understanding is that recovery mode will try for 90 minutes (every 10) and then abort the entire sequence, period.
It is probably true that a large percentage of aborts are due to clouds and most of the rest due to failed attempts at some aspect of an event (flips, guiding, etc) so tend to occur at variable times in a set of events. It is also common for the user to have created several events after the light frames to take calibration frames (darks, flats, maybe bias). In the current SGP an abort after failed recovery, will, unless I am mistaken, just stop, thereby failing to take those frames or perform any actions that might be part of them (such as park mount, close dome/roof, etc). The user would have to do those calibration frames “manually” in the morning, which (because it is light and maybe hot or time to leave) could be problematic.
As part of the recovery mode, allow the user to specify an event number (of the current target) to go to once recovery has failed. If that field is left blank, the current abort procedure would apply. If the specified event number was before the current event (the event during which the abort occurred), it would also use the current procedure. Alternatively one could just have a checkbox that would say “on abort, go to the last event of the current target”. If, by chance, the abort had occurred during that event, it would also abort just as it presently does.
I can see how this could trigger lots of “repeat failures” if the user tried to go to another light frame event instead of darks/flats so one might want to limit this to go to just calibration frame events, especially if mount safety issues might be involved. One probably wants the mount either disconnected or tracking stopped.
Let’s just call this “ATC” or “Abort to Calibration”.
NOTE: One might also have triggering this routine as another option for a manual abort.
This could maybe be combined with mads0100 suggestion to be able to set time/number of recovery mode attempts?