Snapshots are handled in series by the cluster controller

[jackkleeman]

Currently its not useful to submit snapshots concurrently across partitions, as the admin handles them in series. When there are lots of nodes and partition leaders are spread across them, its valuable to do snapshots concurrently. And also in case of large numbers of partitions, each snapshot is likely to be fairly small, so even on a single node we might want to do several at once