Glossary
What is change data capture (CDC)?
Change data capture identifies and delivers only the records that changed since the last sync, instead of copying the whole dataset each time.
Change data capture (CDC) is a set of techniques for detecting which records changed in a source system and passing on only those changes. Instead of re-reading every employee or contact on each run, an integration processes inserts, updates and deletes since the last checkpoint, which is faster, cheaper and kinder to rate limits.
How does it work?
| Method | How it detects changes |
|---|---|
| Log-based | Reads the database transaction log |
| Timestamp / “changed since” | Queries records modified after a checkpoint |
| Events | Receives webhooks or messages per change |
| Snapshot diff | Compares full extracts, as a fallback |
Whatever method you use, keep a durable checkpoint, such as the last processed timestamp or log position, and update it only after changes are applied. Pair CDC with periodic full reconciliation to catch anything the incremental process missed.
Where it shows up in integrations
- HRIS syncs APIs such as BambooHR’s “changed since” endpoint return only employees changed since a date.
- Warehouse loading CDC keeps a data warehouse close to real time.
- File feeds Comparing today’s export with yesterday’s finds changes when no API exists.
Common pitfalls
- Missing deletes, because timestamp queries rarely return removed records.
- Losing the checkpoint after a failure and skipping changes.
- Clock differences between systems causing missed updates at the boundary.