[rabbitmq-discuss] Towards better handling of RabbitMQ connection/channel failures

[Tim Watson]

On 19 Sep 2013, at 01:37, Jonathan Halterman wrote:
> I wrote a test attempting to reproduce what my actual client is experiencing, but I was only able to come close to reproducing my client's results when pushing a lot of volume down to the consumers, and even then it was not consistent enough to draw any conclusions. At this point I'm satisfied to simply tweak my client to account for potential delays in ShutdownListener calls and move on. I just wanted to be sure that there were no mechanisms introduced by amqp-client which could be causing any additional ShutdownListener delays, and it sounds like there are not.
> 

I'm pretty sure there's nothing untoward going on in the java client in that respect. You might want to consider enabling AMQP heartbeats to ensure 'timely' detection of disconnects in situations where there's a possibility of network interruption.

Cheers,
Tim