TL;DR: This incident should now be resolved.
Hi folks - apologies for the silence here. We’ve been hitting capacity on a crucial database this week, in a way that required a brief (seconds) interruption in service to rectify. We were hoping to wait it out until the weekend, or at least a less busy time of day, but as we investigated we discovered that the issues were bad enough that we just had to push the button. We have now done that, and service should be restored.
I apologise for the rough ride using the editor today (the constraint probably bit a couple of other places, but the editor was hit the hardest).
We should now be back on an even keel, and the versioning issues (which were downstream of the same issue) should now be resolved. If you’re seeing any ongoing effects, please report them!