Failed deployment doesn't clean up state

### Description / Background

If Sleeper deployment fails due to a CDK error, then stacks are not cleaned up. Main instance CDK stack was left in failed "Roll back requested" state and the instance "artifacts" stack was left intact and not cleaned up.

### Steps to reproduce

1. Set instance.properties values to something invalid (e.g. set a timeout to longer than 900 seconds, set an EC2 instance type to a non-existant type, etc.)
2. Attempt deployment.
3. CDK deployment will fail.

### Expected behaviour

System should not be left in an inconsistent state that requires manual cleanup.