After seeing the posts yesterday, I think I was able to track down the source and get it fixed -- was working on this roughly 9am - 10am EST yesterday (Sat Jan 7).
Would also have been a short period, only a few minutes long, when other errors might have been generated a couple of hours later.