Skip to main content

Create pipeline

POST 
/platform/pipeline

Create a pipeline with extraction jobs, transformations and output configuration. Use the same connector type for all jobs. JDBC uses load_type: full_load, incremental or incremental_with_qualify; deduplication also needs primary_keys. Singer and S3 examples use replication_method: INCREMENTAL or FULL_TABLE.

Choose a UUID id, customer-scoped unique name, optional description and cron, and jobs. Each job has input, optional transformations and output. Connection-manager authentication uses auth_parameters.auth_type: connection_manager and the config_id from Dadosfera Connections. A null cron represents manual execution.

Contract caveat​

The imported PipelineConfigWithJobs requires customer_id, while source prose says Maestro injects it from authenticated identity. Deployed behavior is unverified; callers must not assume this field allows arbitrary tenant selection. The generic REST source example omits auth_parameters, which the imported JobInput requires. Other source examples include connector-specific fields not individually modeled in the open input/output objects. Confirm those differences before copying them.

The imported responses document 200 for creation, 400 for validation or an existing pipeline, and 409 for an output table-name conflict. Inspect the result before executing the pipeline.

Request​

Responses​

Pipeline created successfully