No need for those journal client to be on kafka nodes.
This is also an attempt to clarify where to deploy our backfill configuration.
In effect, this moves the service from:
- prod: getty -> saatchi
- staging: journal0 -> scheduler0
- deactivate puppet on saatchi, scheduler0
- stop old services on getty and journal0
- (clean up)
- puppet run on saatchi, scheduler0