spryker uses some processes, that are not publicly available I think, to deploy in aws. However, your approach should work for internal usage (we use the docker sdk to provision the project in CI pipeline for example). The docker sdk would only create the endpoints internally. So to access it from the outside, yes, you need to create the subdomains in aws to point to the ec2 instance. And it should be easier, if they are the same as the internal docker/sdk endpoints