Great example and showcase. However, I think it is important to note that the example shown involves forking a new process, which be definition is multiprocessing, and not multithreading. Have you done a comparison overhead between worker_threads and child_process ?