JDBC Source Node
Quick Reference
Connection URL
The JDBC connection URL for your database.
ex: jdbc:postgresql://localhost:5432/mydb
Use Credentials Database username and password credentials.
Query
The SQL query used to read data. Use :watermark as a placeholder when reading in streaming mode.
ex: SELECT id, name FROM events WHERE id > :watermark ORDER BY id
Watermark Column
The column used to track progress in streaming mode. Setting this enables continuous polling.
ex: id
Poll Interval (ms)
How often to re-run the query in streaming mode, in milliseconds. Default: 60000 (1 minute).
Fetch Size
Maximum number of rows returned per fetch. Default: 1000.
Overview
The JDBC Source node reads data from a relational database and passes each row as a record into your workflow. It supports both batch mode (reads all rows once and stops) and streaming mode (continuously polls for new rows based on a column you choose).
Supported databases include PostgreSQL, MySQL, MariaDB, Microsoft SQL Server, and Oracle.
Configuration
| Field | Description | Required | Default |
|---|---|---|---|
| Connection URL | The JDBC URL for your database. Includes the database type, host, port, and database name. | Yes | — |
| Use Credentials | Select or create a Username/Password credential for database authentication. Leave blank if your database does not require a password. | No | — |
| Query | The SQL query to run. In streaming mode, include :watermark as a placeholder — it will be replaced with the last value seen in the watermark column. | Yes | — |
| Watermark Column | The column that tracks how far the query has progressed. Setting this field enables streaming mode. Typically an auto-incrementing ID or a timestamp column. | No | — |
| Poll Interval (ms) | How often to re-run the query in streaming mode, in milliseconds. For example, 5000 checks every 5 seconds. | No | 60000 |
| Fetch Size | How many rows to retrieve at a time from the database. Reduce this if you run into memory issues with large result sets. | No | 1000 |
Batch vs. Streaming Mode
Batch Mode
Leave Watermark Column empty to run in batch mode. The query runs once, all matching rows are read, and the workflow completes. Use this for one-time data loads or scheduled exports.
Streaming Mode
Set a Watermark Column to enable streaming mode. The node polls the database repeatedly on the configured interval. Each run fetches only rows where the watermark column value is higher than the last seen value, so no rows are duplicated.
Your query must include :watermark as a filter and must order results by the watermark column:
SELECT id, name, value FROM events WHERE id > :watermark ORDER BY id
For timestamp-based progress tracking:
SELECT * FROM logs WHERE updated_at > :watermark ORDER BY updated_at
Related Nodes
- JDBC Sink: Write records to a relational database
- Kafka Source: Consume streaming data from Kafka