JustAnotherWriter, that’s just about the size of it. From the usage angle there are no good useful patterns. Even the new vs old device and processor type clues, fruitful and interesting though they seemed at first, ended up being more indicative of averages that absolutes.
But the log files that some of you have taken the time to send in have at least shown some more concreteness in their similarities. They do very much seem to indicate a RAM issue that, while it only happens in some usage scenarios, at least cause very similar logging results. The question now is how to optimise this main sync loop, if that proves to be the problem, particularly if we can’t get the fault to happen in the first place. Could be general optimisation is good enough, we’ll have to see.
My very low-info guess at this point is that iOS 13 has tighter restrictions on per-process RAM usage, and that some combination of unknown factors causes that to become more prevalent.