○ We will call the application that’s already up and running the “blue” instance. ○ We would then need to create an identical “green” instance running in parallel. This is pretty easy to do if all the infrastructure is managed in Terraform. Can be as simple as changing a few variable names! ○ Once a new version of the application is ready for testing, we could deploy it to the “green” instance and then slowly shift traffic over to the “green” instance using a Route 53 weighted routing policy. ○ Once fully-satisfied with “green’s” performance and stability, we could perform a full cut-over to the “green” instance and stop routing traffic to “blue.” ○ If wanting to save costs, it may be worth it to tear down the “blue” instance as a cleanup step. But it’s always a good idea to keep it up just in case there is a problem with “green” and we can immediately route traffic to the stable “blue” instance.