A nice feature would be to log errors and/or possibly drop events if the ASIO queue is being overwhelmed. ASIO has no direct way to monitor the task queue, but we could have a 'heartbeat' task to time the responsiveness.