We are back up now. We rolled out some changes this weekend that didn't work properly under high load, causing the audio server #2 to go into a hung status. Unfortuantely, it happened during the evening and our monitoring systems didn't detect a problem since the server was still running, just not serving feeds.
We've rolled back the change and now will investigate why this became an issue yesterday and how can prevent it in the future.
Sorry for the outage and the problems.