Getting started with Change Data Capture in Amazon Aurora DSQL

This article introduces Change Data Capture (CDC) in public preview for Amazon Aurora DSQL, enabling near real-time streaming of database changes to Amazon Kinesis Data Streams.

• CDC captures row-level INSERT, UPDATE, DELETE operations and streams them to external systems
• Aurora DSQL CDC operates at cluster level, capturing changes across all tables without selective filtering
• Events include metadata: operation type, timestamps, transaction ID, schema, table, and row data
• INSERT and UPDATE operations represented as "c" events; DELETE operations as "d" events
• Events delivered unordered; downstream consumers must handle duplicates and out-of-order processing
• Configuration requires IAM role with Kinesis write permissions and trust policy for Aurora DSQL
• Supports real-time analytics, event-driven architectures, data replication, and audit trails
• Maximum row size 2 MiB; configure Kinesis record size accordingly to prevent delivery failures
• Best practices: use on-demand Kinesis capacity, implement idempotent processing, handle duplicates

Aurora DSQL CDC enables building real-time data pipelines without custom replication solutions, supporting scalable event-driven systems and analytics.