Accounting for regional failover
It is not a matter of if, but when, a given cloud provider will experience a news-worthy regional disruption. In my experience, this happens about every 2-3 years. Another major event occurred just 2 months before I started writing this chapter. It was a little overdue, but it happened nonetheless. It is inevitable.
Designing a system to run across multiple regions may seem daunting at first, but it is crucial because we rely on a cloud provider's serverless features as our platform. Fortunately, deploying serverless features in multiple regions is neither difficult nor expensive. In Chapter 9, Choreographing Deployment and Delivery, we will see how to deploy to multiple regions and leverage one region for canary deployments.
And, as it turns out, multi-regional deployments are not daunting. We are already designing for regional failover because we are designing for eventual consistency. And eventual consistency extends cleanly to support...