Feature: endless asyncio tasks hypervisor

[Lancetnik]

Now we have some endless asyncio.Tasks consuming messages like here or in kafka/confluent/redis as well.

They have no chance to fail, but anything can sometimes things happens and users even don't see any exception about it. Thus, we should log this tasks failure and try to restart them.

• pass broker.logger to subscriber through setup method
• add task done callback to log an exception and restart it

It should helps users to diagnostic the problem and create Issue about accident and improve the system susteinability