It looks like we've got a handle on the backend failures issue now; identifying and removing a very small bit'o'code, and monitoring for a few hours, gives us confidence this issue has been resolved.
I thank you all for sharing your data points and especially offer thanks to those of you who remain positive as we work to resolve these things, remembering that on the other side of every message posted, there are actual living, breathing people reading those messages...who, at the end of the day, are all just trying to do their jobs as best as possible with the tools they've been given.
Regards,
WUToddG