background_distributed_schedule_pool_size setting doesn't do what it says

[ivan-tkatchev]

Looking at the source code and observing real-world behaviour, I see that asynchronous distributed inserts are actually single-threaded. (Or, rather, roughly one thread per table.)

Is this expected behavior? If yes, then perhaps this should be documented, because the current documentation makes it seem like increasing background_distributed_schedule_pool_size actually does something.

Clickhouse version 23.11.1.1