Why this matters
A description written in Collate is invisible to an engineer working in the warehouse or a tool reading table comments. Reverse metadata writes descriptions and tags back to the source system so those readers see them too.
Before you start
Collate stops being a read only mirror
Ordinary ingestion is one directional: the source is authoritative and the platform reflects it. Reverse metadata inverts that for the fields humans curate, letting descriptions and tags authored in Collate land in the source system where other tools can read them.
Two way sync needs a rule about who wins
Once metadata can move in both directions, conflict is possible. Someone edits a comment in the warehouse while someone else edits the description in Collate. Knowing which side is authoritative for which field, and on what schedule the write happens, is what keeps the two from fighting.
Run history is the audit trail
Because this writes to production systems, the logs matter more than usual. Run history tells you what was written, when, and whether it succeeded, which is the evidence you need when someone asks why a comment in the warehouse changed.
The interactive demo
Work through the demo below. It runs in your browser with nothing to install. Watch for:
- Configuring reverse metadata
- Configuring run schedules
- Examining run history and logs
This demo is also available on its own at the Learning Center.
After this module you should be able to
- Configure metadata to flow from Collate back to the source
- Reason about which side is authoritative for a given field
- Use run history to verify and audit what was written