Restarting the entire system is not a problem in our use-cases, but it is frustrating that receivers need to be stopped when almost all code updates do not affect them.
There is also an argument that decoupling the receivers from the rest of the system would make it easier to deploy in security-constrained environments. For example, if the receivers could run in a DMZ and the rest of the system could retrieve messages by calling out to the DMZ it would provide a more secure deployment.
This would perhaps require a 2nd MQ - one running for the receivers and one running for the rest of the system. Alternatively, FTP or scp could be used to get messages.
Restarting the entire system is not a problem in our use-cases, but it is frustrating that receivers need to be stopped when almost all code updates do not affect them.
There is also an argument that decoupling the receivers from the rest of the system would make it easier to deploy in security-constrained environments. For example, if the receivers could run in a DMZ and the rest of the system could retrieve messages by calling out to the DMZ it would provide a more secure deployment.
This would perhaps require a 2nd MQ - one running for the receivers and one running for the rest of the system. Alternatively, FTP or scp could be used to get messages.