Improve workflow speed

[qqmyers]

From DANS' work to migrate ~100K datasets using a pre-publication workflow, @janvanmansum discovered that the 1 second delay added prior to the workflow start (which was part of changes needed to make post-publication workflows like the archiving workflows succeed) was a) adding ~27 hours to their runtime and b) wasn't necessary for pre-publication workflows (see GlobalDataverseCommunityConsortium#4). Building on that, a solution for both just needs a flag that indicates whether it is safe to find() the dataset from the database or if a delay is needed. A PR to follow - with more explanatory comments.