We are looking at metrics generated from New Relic on our application. It is fairly high throughput sitting at around 5k RPM. We notice that a small percentage of requests, ~1% get stuck inside "HttpServlet30Dispatcher.service()" which is the jboss impl of resteasy. For the requests that get stuck there, it looks like HttpServlet30Dispatcher (or some code underneath it) consumes 100% of the time of that request. We see requests that normally take 50ms go up to 20s+.
Unfortunately I can dig into HttpServlet30Dispatcher code very easily, but it definitely doesn't get to our Application code which I can profile, I can see it is only 0.1% of those queries.
Does anyone know what could cause this? It feels like WF could be blocking the request waiting for a free thread or something. Can anyone confirm, or any advice on how to profile this?