Until now, the HA was focused at the container level.
We have built mechanism to be sure that message are properly send/received by container, and the fail over on that.
But no mechanism has been build on the assurance on the service has been properly treated or not, such as:
-the component destination is really and still a valid destination
-the component destination receive the message but never answer...failure on the message process not handled or hanging of the component...
To deal with these issues, mechanism must be implemented in the router.