I have a simple Spring Boot REST server which I intend to use for load testing modules under development.
Under moderate loads, the clients were experiencing prolonged delays, even timeouts. It turns out that the server was only processing 2 requests at a time.
This has nothing to do with server.tomcat.max-threads. The worker thread pool (http-nio-< port>-exec-< number>) never exceeded 10 when I increased max-threads above 10, but shrank to 1 when set to 1. I assume because the pool did not receive more than the two requests at a time.
What could be limiting the number of concurrent requests to 2 ?