So far nothing conclusive at the software and hardware level, we think it may have been a temporary network connectivity issue at the datacenter level, as it didn't just affect one service.