I don't use it this way. When I want to deploy something I just git push on a branch and CodePipeline take care of the rest. So I don't want to give you wrong information. But it seems that eb deploy is made for that. Be careful, there is two CLIs: - aws [command] [subcommand] that take care of the low level actions (creating EC2 instances, Load Balancers…) - eb [command] that take care of ElasticBeanstalk only, for higher level tasks (deploy, logs, environnement variables) http://docs.aws.amazon.com/elasticbeanstalk/latest/dg/eb-cli3.html

My setup is similar to the above, but I'm using CircleCI to do the build/deploy step instead of CodePipeline

The build step is basically lein uberjar and then creating an ElasticBeanStalk "application bundle" - a zip file with the uberjar in it and some extra AWS stuff like a Procfile. Then the deploy step is basically eb deploy.

You can probably get away with a much simpler setup, but we needed to tweak some things for ~reasons

But I agree that Java SE ElasticBeanStalk is probably one of the simpler and most dev-friendly setups you can work with on AWS at the moment