I'm looking into optimal ways of accessing the HTTP request and response bodies for tracing in a Spring reactive application.
For previous versions, we've leveraged Servlet filters and Servlet request wrappers to consume the incoming request's input stream and hold a copy of it for asynchronous processing of the traces (we send them to Elasticsearch).
But for a Spring reactive app (using webflux), I'm wondering what'd be the most appropriate way to access the requests before they're decoded. Any thoughts?