This way we can give clusters semantic names like "Production" and point the scheduler at those names instead of relying on the job flow ID which is not very human friendly / memorable. Change-Id: Iebeab022e09047b625f4725300d357905555ef04