You adjusted the defaults.
The prior defaults seemed not logical to me .. but did you test what effect those configs have with e.g. different load?

Might be worth to add a little documentation for those configurations.

Prior defaults were not working at all as those properties should be inside "thread-pool-executor" object. If putted in such a threadpool with int.max number of threads was being created.
I did tested with different load and was able to publish more than ~200 m/s an endpoint that responds after 1.5s only by changing the dispatcher settings and parallelism of the connection. Before changing the dispatcher config there were "There are too many in-flight requests." (handled by parallelism) and "Dropped message as result of backpressure strategy!" (handled by tweaking the dispatcher). Maybe the default factor can be left at 3 but i don't see a drawback if set to 4. Nevertheless i don't have a strong opinion on the default value as i will be overriding those later.

Ok.
If you already know you want/need to overwrite them, you could give a hint in the config documentation about how and when to tweak them (e.g. when inspecting certain dispatcher metrics).

But I assume this is highly dependent on the use-case :D