What happens to your web app when a databases update fails? Or the email service is unreachable due to network hiccups? Will the front-end survive bad code crashing the back-end? Admit it - the web app will crash and show a useless "server error" message. But there is hope! Messaging servers, like RabbitMQ, work wonders in the face of failure - decoupling processes and allowing for re-try when things come back up.
In this talk, I'll show you how I re-architected SignalLeaf to deal with crashes and sometimes-connected services. You'll hear tales from the trenches of success and failure with RabbitMQ and Node.js. You'll see how a message queue can help you build a more resilliant web app. And you'll clean up your code at the same time! Rejoice as your application architecture improves by leaps and bounds, with RabbitMQ!
The last paragraph can be dropped. And the narrative of the first two (which is good and interesting) can be made into one paragraph. I'd suggest less specific detail (2008, etc), and more generalities.