Things generally die down in production on Wednesday mornings and Friday nights, in terms of general traffic and usage on the Manager.
The next release of the manager overhauls a lot of the API (i.e. back-end) functionality and is going to require a brief downtime, so that's going to happen tonight when the user active user count gets down into the low double- or single-digits.
The downtime should only be momentary (but could go on for a few minutes, especially if I screw it up as badly as I screwed up the last one).
All of which is to apologize in advance for any issues tonight and say thanks again for using the Manager!