It looks like WaitForRDSRRStack1 does not have a path to DeployFailedState1 (the WaitForRDSStack1 uses RDSStackCompleteChoice to do this), therefore if the Wait Lambda times-out, it moves on the DeployDMS, which then fails when it tries to retrieve the DB Server name here from the RR stack outputs (which do not yet exist).
Redriving the Lambda (or presumably the entire workflow) once the RR stack is complete, fixes this
It looks like WaitForRDSRRStack1 does not have a path to DeployFailedState1 (the WaitForRDSStack1 uses RDSStackCompleteChoice to do this), therefore if the Wait Lambda times-out, it moves on the DeployDMS, which then fails when it tries to retrieve the DB Server name here from the RR stack outputs (which do not yet exist).
Redriving the Lambda (or presumably the entire workflow) once the RR stack is complete, fixes this