Currently we have to SSH into each node and run docker system prune to clean up old images / data. This seems fairly impractical for large swarms. It would be great to have command that can be run from the leader that instructs all the other nodes to clean up old data.