Import
The Import module provides a managed ingestion workflow for bringing external data into a VeloDB Cloud warehouse. Use it when you want VeloDB Cloud to create and operate ingestion jobs through a visual wizard.
Import is suitable for the following scenarios:
- Loading existing data from a source system into VeloDB Cloud.
- Continuously ingesting changes from MySQL or PostgreSQL through CDC.
- Consuming records from Kafka-compatible event streams.
- Loading files from object storage on a schedule or as a batch job.
If you want to query external data in place without moving it into VeloDB Cloud, use Add Catalog instead. If you are moving a full Apache Doris or VeloDB warehouse, use Migration.
Supported sources
| Source | Use case | Guide |
|---|---|---|
| MySQL | Initial load plus incremental CDC, or incremental-only ingestion | MySQL |
| PostgreSQL | Initial load plus incremental CDC, or incremental-only ingestion | PostgreSQL |
| Kafka-compatible streams | Continuous ingestion from Kafka topics, including Apache Kafka, Confluent Cloud, and Amazon MSK | Kafka |
| Amazon S3 | Batch loading files into VeloDB tables | S3 |
Before you begin
Before creating an import job, make sure you have:
- A running VeloDB Cloud warehouse.
- Source connection information, such as host, port, database, username, password, access key, bucket path, or Kafka bootstrap servers.
- Network access from VeloDB Cloud to the source system. Start with Connect for Import.
- If the source is in a private VPC, configure private connectivity. For SaaS warehouses, see Grant VeloDB access to your VPC.
- Source-side permissions required for the selected ingestion mode.
- A target database in VeloDB Cloud, or permission to create one during the wizard.
Keep Amazon RDS and Aurora private connections available
If you import from Amazon RDS or Aurora through an AWS Network Load Balancer (NLB) with IP targets, keep the target group synchronized with the database endpoint. Failover, maintenance, or scaling can change the endpoint's private IP addresses and leave the NLB pointing to stale targets.
See Automatically Sync IP Targets for Amazon RDS and Aurora Endpoints to configure AWS Lambda and Amazon EventBridge for recurring synchronization and verify failover handling.
Next steps
Choose the source-specific guide for your import job: