on_disconnected occasionally not triggered

[sein-park-development]

Hi, is there any scenario that on_disconnected won't be triggered, in intermittent manner?

My services on AWS ECS (docker based service) does not always trigger on_disconnected when connected mq component dies. Only 70% of the services are receiving the trigger so I don't get any clue what the proportion means.

[jasonrbriggs]

Haven't seen that problem, but I don't have AWS to test with either. It could be a bug, but hard to know for sure. I'd suggest setting up a docker instance off AWS and create a test to see if you can replicate the problem.

[CorvetteCole]

I have been seeing this problem lately. In fact, I can reproduce it 100% of the time with the following:

• RabbitMQ set up on remote server
• TLS enabled
• Acking enabled
• Heartbeats enabled

Further details can be of course provided, I may file my own separate issue soon. If I start this and then unplug the ethernet cable connected to the server, I will get a print message about the heartbeat receive diff but then nothing - it just crashes. No further messages even in debug mode. Currently digging further in to this

[grundleborg]

@CorvetteCole did you ever find any further information about this. I'm seeing the exact same behaviour as you.

[CorvetteCole]

@grundleborg Unfortunately not, and I've since left the company I worked for that I was debugging this stack for so I don't think I'll be further pursuing it. I wish you the best of luck though! I ended up implementing a workaround by way of an external thread that was manually checking status and triggering the on_disconnect function