Does Wildfly IO Subsystem implements something like the "acceptCount" found on the Tomcat HTTP Connector ?
"The maximum queue length for incoming connection requests when all possible request processing threads are in use. Any requests received when the queue is full will be refused. The default value is 100." ref: 
Some kind of Queue threads limit. What happens when the "ask-max-threads" and "io-threads" limits are reached? I don't know if this concept applies to the (X)NIO implementation used by Wildfly.
Is there a way to monitor those limits in order to know if the queue is being used?
I've found a discussion  on another forum, but I'm not sure if this still applies to the latest Wildfly version configuration.