Hi Team, Do the upstream and downstream datasets need to be specified manually, or will it take automatically when we connect and load the database schema to datahub?

It depends on the source if it can capture upstream/downstream lineage automatically. Which source do you want to use?