Default queue closes abruptly

[bevzzz]

I have a simple Go app with producer and consumer processes communicating via Redis that runs in a Docker container.
The server does not define any queues explicitly, so the default one is used.

Trouble is, asynq seems to sometimes delete/close the default queue, even though it often has tasks scheduled in it. I tried setting LogLevel to DEBUG and checking the logs from the Redis container, but there's little useful I found. When the queue is removed, all I see happening is:

• "asynq" CLI suddenly shows an error message saying queue "default" does not exist for the queue that existed just a moment ago
• My asynq.Server pulls tasks but gets a "All queues are empty" error all the time, so the workers do not pickup new tasks.

Could it be because the connection to the Redis container is getting lost?
If so, how can I recover from it?

What's interesting is that on enqueuing a new task, the "default" queue is created again and the new task is being picked up by workers.

My server setup is rather simple:

srv := asynq.NewServer(
	asynq.RedisClientOpt{Addr: redditAddr},
	asynq.Config{
		Concurrency: 3,
		LogLevel:    asynq.DebugLevel,
	},
)

Am I missing something?