I have a Webflux application, which accepts incoming json requests and returns flux responses. Each request in turn initiates a request to a another micro-service using Webclient.
How many concurrent requests can such a webflux application handle, and how many concurrent requests can such a WebClient make? More importantly are there any reactor/netty/webflux configuration parameters that help in deciding the number of concurrent requests that can be processed.
The application is mostly I/O bound, and almost all of the latency is from the outgoing requests via webclient. It's a quad-core machine with 16GB RAM.
(edited) - I am basically looking to understand how many threads are generated, the threading model in such an application, and configuration parameters that Netty provides. Note - I am not looking to actually get throughput/latency numbers for my setup (I can, and have already load tested the application).