At approximately 7:45pm Central, there was a brief service interruption affecting approximately 20% of users, resulting in some users being temporarily directed to different versions of the same room. The issue was quickly localized as due to a single server experiencing instability in the core authentication service. Services on that server were restarted, and the issue resolved itself. Additional capacity was also brought online to enable rapidly moving that subset of users to a different server if the problem recurs. Further investigation into the cause of the instability is ongoing, but we do not expect additional issues at this time.