What is Change Data Capture (CDC)? Real-Time Database Sync Explained
Change Data Capture (CDC) captures every row-level change in a database — insert, update, delete — in real time using the transaction log. Learn how CDC works, how it differs from polling, and where it powers streaming data pipelines.
What is CDC?
Change Data Capture (CDC) is a technique for tracking and capturing every row-level change — inserts, updates, and deletes — made to a database, in real time, by reading the database's transaction log. Rather than querying the database periodically (polling) and comparing results, CDC reads the write-ahead log (WAL) or binlog directly and streams each change event as it is committed. This enables low-latency data replication, event-driven architectures, and audit trails with zero additional load on the source database.
- Captures every insert, update, and delete at the row level in real time
- Reads from the database transaction log — no polling, no extra query load
- Enables streaming pipelines with sub-second latency
- Preserves the full change history including deletes — which polling misses
- Common CDC tools: Debezium, AWS DMS, Fivetran, Airbyte
How CDC Works
CDC connectors attach to the database's replication stream — the PostgreSQL WAL, MySQL binlog, SQL Server CDC tables, or Oracle LogMiner. Each committed transaction produces change events with the before and after state of the row. These events are published to a message queue (Kafka) or consumed directly by a downstream pipeline.
Enable replication on the source
Configure the database to expose its transaction log for replication (e.g., set wal_level=logical in PostgreSQL, enable binlog in MySQL).
Connect the CDC connector
A CDC tool like Debezium subscribes to the replication slot or binlog and converts raw log entries into structured change events (before/after row state, operation type, timestamp).
Stream change events
Events are published to a message queue (typically Kafka) as they arrive, enabling downstream consumers to react in near real time.
Consume and apply
Downstream systems — data warehouses, caches, search indexes, microservices — consume the event stream and apply the changes to stay in sync with the source.
Common CDC Use Cases
Real-time data warehouse sync
CDC streams every transaction from a production OLTP database into Snowflake or BigQuery in seconds — eliminating the 12-hour gap of nightly ETL batch jobs.
Cache invalidation
When a product price changes in the database, a CDC event triggers immediate cache invalidation in Redis — ensuring users never see stale prices.
Microservice event sourcing
Rather than coupling microservices directly, CDC turns database changes into events on Kafka. Each service subscribes to the relevant events and updates its own read model.
CDC in FlowGenX AI
FlowGenX can listen to CDC event streams as workflow triggers — using Kafka or direct database connectors. When a CDC event fires (a new order row, an updated customer record, a deleted ticket), FlowGenX launches the corresponding agentic workflow in real time, with the full before/after row payload available as context for AI decision-making.
Frequently Asked Questions
Related Terms