This is a follow-up from ENTMQBR-1133. A Spring Boot application that uses the Artemis Core-JMS client to connected to AMQ 7 does not respond correctly to transport failures. When ENTMQBR-1133 was raised, we were not sure exactly what the client application was doing, so it wasn't clear where the defective behaviour was. Now, however, it seems that the client is not calling registered exception listeners, even in circumstances in which a messaging operation fails with an exception.

Since the Spring Boot infrastructure is listening for exceptions, and responds by recreating its connection pool, the effect of not calling registered exception listeners is that the application continues to use stale connections.

I have attached unit tests.