The idea is:
allow customizing of number of core threads per main/worker to exceed 1.
My use case:
queue mode is great but does not sit well with edge architectures (6-10 low power cores) where main is constantly a bottleneck due to sequential RAG flows, and multi-main is not really a solution (correct me if I am wrong) since the two mains would not split the work of a single workflow.
I think it would be beneficial to add this because:
it would allow both on-prem flexibility, and it would become possible for cloud to scale up cores within the same instance and not just add or remove workers, balancing multiple use cases (add workers for concurrency, cpus for horsepower) making N8N faster with a lower energy footprint.
Any resources to support this?
I base this feature request on the assumption that if main is constantly hitting between 99% and 101% , it is in fact a bottleneck.
Are you willing to work on this?
I don’t have the skills to do that 
my understanding is that the actual execution of the code of a single workflow would need to be split with certain type of operations done by one core and others done by another core