Pete : I think you misread my patch. Concurrent calls don't get parallelized. I didn't really understand what this concurrent request timeout really represents, but it seems awkward to me that a request has to fail because an existing request is already being processed. I changed the behaviour from "wait for the timeout" to "wait for the current action to finish". You could still say : "ok let's increase the request timeout".... but you're just postponing the problem because if you take a stack of conversation events to be processed, for them to be processed correctly they need to be all executed within a shorter time than the timeout. View the original post : http://www.jboss.com/index.html?module=bb&op=viewtopic&p=4080602#... Reply to the post : http://www.jboss.com/index.html?module=bb&op=posting&mode=reply&a...