Planning for failure is a lot more challenging than what happens when an entire AWS Zone or Region goes dark.
What about handling high level dependencies of your application? What even are your application's dependencies?
We'll use the AWS Kinesis outage as an example to understand how to monitor your application and deal with partial failures. Learn the basics of how to measure a feature or dependency of your application, step through the challenges of being able to manually fail over, and plan for being able to automate that fail over process.