Configuration
An ingestion configuration defines how source data is interpreted during ingestion.
It references the data schema used to validate the source columns. Ingestion jobs then reuse the configuration when creating or updating datasources.
Create a configuration
Create an ingestion configuration after you have created or selected a data schema.
Save the returned configuration id. You will use it when starting ingestion jobs.
Observation fields
During ingestion, the platform can add observation fields that help identify, track, and process rows.
By default, these fields use standard names such as observation_type, observation_source, and observation_increment_index. You only need to configure observation_specification when your source data already uses different column names for these concepts.
Training and test rows
If your source data includes a column that separates training rows from test rows, map that column through observation_specification.observation_type.
If no observation type is provided, ingested rows are treated as training data.
Retrieve configurations
Retrieve ingestion configurations when you need to reuse an existing configuration.
To inspect a specific configuration, retrieve it by id.
Update a configuration
Update an ingestion configuration when you need to rename it or point it to a different data schema.
Schema changes apply to future ingestion jobs that use the configuration. Existing datasources keep the schema that was applied when they were ingested.
Delete a configuration
Delete an ingestion configuration when it should no longer be used by future ingestion jobs.

