No. You can do that much easier in docker swarm.

In my case I just take the 5-10 second downtime in between container restarts.