Given throughput/latency targets, size CloudHub workers (vCore sizing) and worker count, and select horizontal vs. vertical scaling, parallelization, and caching to meet capacity goals.