I don’t understand the benchmark. Different libs have different worker counts. Is the actual throughput test doing any async work? Like calling a db or a REST call before it serves a response?

How is this an apples to apples benchmark ?