Data Sources
A data source connects a physical model in a Kubling VDB to data that Kubling can query or mutate. The adapter boundary determines where connectivity, metadata discovery and source-specific execution take place.
Kubling 26.4 supports the following main categories:
- gRPC Provider: An external process implements the public provider contract. The provider owns native connectivity, credentials and source-specific execution, while Kubling handles federation and planning. Start with the Providers overview.
- Database: Built-in relational database connectors import remote metadata or use explicitly defined foreign tables. See Database data sources.
- Script Document: JavaScript handlers expose semi-structured documents or APIs as tables. See Script Document.
- Embedded and Portable Local: Local storage is hosted with the engine for lightweight or self-contained models.
- Industrial: Built-in protocol adapters integrate systems such as SNMP and Modbus.
- Proxy: Composite and routing data sources operate over other Kubling data sources instead of connecting directly to a remote origin.
The built-in KUBERNETES, REDIS and CASSANDRA types were removed in
Kubling 26.4 and can no longer be used. Migrate existing VDBs to their
official external providers by following the
migration guide.
Documents
Some data sources exchange JSON or other semi-structured objects. Throughout the documentation, these values are called documents regardless of whether the remote system is a document database, an API or another kind of source.
Kubling can retain documents as JSON columns, expose selected fields as typed relational columns and define synthetic tables over nested structures when the data source model supports it.
For the role of data sources inside the engine, see the architecture overview.