Replication uses Postgres logical replication to replicate database changes to external destinations like Analytics Buckets and BigQuery. Changes are captured from the Write Ahead Log and delivered in near real-time to analytical systems.
Key benefits
- Near real-time sync: Changes replicated as they occur using WAL reading.
- Analytics Buckets support: Append-only changelog format in Iceberg.
- BigQuery integration: Direct replication to Google's data warehouse.
- Complete change capture: INSERT, UPDATE, DELETE, and TRUNCATE operations.
- Managed pipeline: Monitor status, lag, and errors in dashboard.
Destinations
Analytics Buckets create append-only changelog with cdc_operation column, preserving complete change history in Iceberg format. BigQuery creates views backed by versioned tables for efficient querying.
Setup
Create Postgres publication for tables to replicate. Add destination in Replication section of dashboard. Configure destination-specific settings. Monitor pipeline in dashboard.
Requirements
Tables must have primary keys. Logical replication must be enabled.
Replication is valuable for:
- Real-time data warehousing
- Analytics separation from production
- Historical data archival
- Multi-destination data sync
- Compliance and audit trails
Limitations
No DDL support yet (ALTER TABLE, ADD COLUMN). Destination-specific constraints may apply.
Replication provides the real-time data pipeline required for modern analytics architectures.