More information on how to define an etl in a system-under-test can be found there.
The process for testing that an etl run successfully (no failure during the execution, the result of the execution is a success) is really straightforward: you run your etl and check the state at the end of the execution. If the package returns a valid status then the test is successful else it fails. To achieve this, in addition of the etl defined in the system-under-test you will need to define an xml element named successful within the assertion.
A full sample would be:
To assert that a run is not successful just add a not attribute to the successful element.