Your back-end processing supports multi-threading and you plan to leverage this capability in the context...

Each distributed queue member receives as many messages as it dispatches (nothing needs to be done from a RVCMQ standpoint as it is a matter of the multi-threaded message dispatching).

Each distributed queue member receives as many messages as specified by the worker tasks parameter (the worker tasks parameter of each queue member must be equal to or greater than 5).

Each distributed queue member receives as many messages as specified by the scheduler configuration (the scheduler tasks parameter must be set to 5 or greater).

RVCMQ scales by multiple program instances (horizontally), not by multi-threading (vertically).