I've never been completely clear about how the restart (after clouds) works.
It definitely does restart but I don't know exactly how it decides to do this. Maybe there's another setting someplace. I think it just keeps trying to find guide stars and then eventually if guide stars reappear it just resumes where it left off.
There's a current discussion on the INDI Forum (https://indilib.org/forum.html) about fixing a bug that fails to recenter the target (using plate solving) upon restarting after a guide failure. My mount tracks pretty well without guiding, so I am kind of protected from this. But if it's been cloudy for hours my target will be significantly off-center. Hopefully they get that fixed in the next revision.
There are some other quirks about the scheduler that took me a while to understand, but in general I've been really happy with it and with Ekos in particular.
You should certainly visit the INDI forum though -- loads of useful stuff over there (as well as the usual drivel)!