The ec2-plugin currently allow to spawn on demand nodes when the load is increasing and to stop or terminate them when not needed anymore. These instances are spawned by providing an AMI in the cloud node feature.

In the case that you would like to use existing EC2 instances as regular nodes (not through cloud nodes) there is no option to delegate the ec2 start/shutdown of these instances to the Jenkins master.

Adding a new launch method "ec2 controlled" with the ID of the instance(s) as a parameter would allow us to get rid of the AMI config and only provide existing IDs as Ec2 slave nodes.

This is useful when you have instances with lots of enterprise tools config on it which can't be automatically integrated in the enterprise global installed base for example.

Basically this feature is just to add a launch method that would be able to start or stop ec2 instance when required or idle (instead of just stopping the jenkins agent process and let the instance in a running state).