I published Archetype for AWS Elastic Beanstalk, which uses DropWizard instead of Spark (but both use Jetty). It should be pretty easy to adapt it to Spark. Please note that it uses the Docker Solution Stack (more flexible and easily debugged locally). This command should start:
$ mvn archetype:generate -Dfilter=elasticbeanstalk-docker
In addition, AWS allows you to create a custom Java application, such as @ k.liakos, said in his answer.
source share