Context: A Rails app running action cable for websockets using Redis, on an EC2 instance, about to be upgraded to work on multiple instances behind an AWS Application Load Balancer. I've read a few blogs explaining people's setup, but there's a key concept I'm not getting.
Question: My understanding of websockets is that they require a fixed connection between the browser and the server. AWS Application Load Balancer (ALB) handles this by fixing a websocket connection to a specific instance behind the ALB. But suppose someone is subscribed to a channel and has a connection through the ALB to server A, and an event happens on server B which triggers a websocket broadcast for that channel.
How do the clients connected to that channel on server A ever get the message?