JmsMessageSenderIntegrationTests is still flaky

[snicoll]

We fixed the issue #1539 but the change introduced another problem. It looks like we should 'wait" for the request/reply to happen before attempting to read the reply.

[snicoll]

I've pushed something to browse the response queue and wait for a message. Let's see if that's helpful. No idea why temporary queues aren't affected though.

[snicoll]

We're getting there. I managed to reproduce a 3s timeout for the thing that's waiting for the reply so it might be some load on a runner that makes it unusually slow