We noticed some irregular behavior with the profile database during the day. We believed that it eventually would have started to affect user saves negatively. In order to circumvent this, we decided to take down the service, restart the database, and run some stability tests to ensure that intermittent and unplanned outages didn't occur over the weekend. We would rather control the problem vs. let it cause chaos at an unknown time. This decision was made around the time the live stream was starting. As for the maintenance itself, the irregular activity was resolved and our stability tests looked good. We are now working with our vendor to prevent these types of issues in the future. Thanks to everyone for their patience with this matter.