Schedules and runs
A flow you can only run by hand is a script with a nicer interface. These are the parts that make it something you can rely on.
Schedules
Put a flow on a schedule and it runs without you. If the computer was asleep or the app was closed when a run was due, the flow catches up on what it missed rather than silently skipping it.
Schedules are part of the Plus and Pro plans.
Incremental runs
A second run should move what is new, not everything again. Two mechanisms do that:
- Cursors, for sources with a natural order (a database table read by keyset paging).
- Content hashing, for sources without one, so a file that has not changed is not reprocessed and a file that has changed is.
The run ledger
Every run records what it read, what it wrote, what changed, and what it skipped. When someone asks why a number moved, the answer is in the ledger rather than in your memory.
Drift detection
If the shape of the source changes underneath a saved flow, for example a column disappears or changes type, the flow notices and tells you, instead of writing wrong data into a table that still has the old shape.