[bjslade]

This is the approach we’ve taken with other similar systems too. Things can go so horribly wrong underneath your containers that the concept of having only one cluster in a prod scenario and maintaining it mid-air would be unthinkable.

Also agree on keeping state outside. Maybe the relevant tech will be mature sometime soon but we’ve seen orchestration bugs do nasty things to stateless containers that would have been a nightmare if state had been involved.