The heart of any service is its availability. Even the coolest game, mobile app, or website is hard to love if users can't rely on it to be available when they need it the most. In this session, we dive deep into how we think about availability of services at AWS and the measures we take to maximize our availability at scale: starting from multi-datacenter services to our deployment methodology, canaries, and fault tolerance in general. In this session, you will get a comprehensive view of how different organizations think about availability and the tradeoffs we make on our way to building highly available services.