Availability Zones

WHAT IS IT? [Azure AZ/ Google AZ]
Each AWS region has multiple availability zones (typically 3 or more), which contain several subnets. You can and should utilise multiple availability zones (or AZs) to create redundancy, allowing for high-availability and fault tolerance. An AZ consists of one or more AWS data centres. When you create a VPC, it spans all availability zones across a region, and you can add subnets to each AZ.
Availability Zones exist so AWS can ensure that its customers’ applications are consistently available in the face of issues with the servers themselves. That could come in the form of component failure or a backhoe cutting the network link to a data centre. AZs are situated geographically far enough that a localised problem shouldn’t affect more than one at a time.

WHAT’S THE FUSS?

You can effectively use AZs to create high availability & fault tolerance for your applications. HA means that you’re minimising downtime, while fault tolerance means you can continue serving services even if things start to break down. By launching instances in multiple AZs, you can protect your service from going down.
Say we have 2 AZs, with instances in public subnets in each AZ. You can configure your EC2 instance in 1 AZ to automatically replicate over to another EC2 instance in another AZ in case of a failover scenario - but you still have to ensure the backups in the first place!