Skip to main content
Pipeline control components manage the flow of execution through an orchestration pipeline. Rather than moving or transforming data themselves, they determine which components run, in what order, and under what conditions.

​
Use cases

Pipeline control components are especially useful for building pipelines that need to handle errors gracefully, avoid duplicating logic across projects, or coordinate several dependent pipelines as a single workflow. Use pipeline control components to:
  • Branch pipeline logic based on conditions, running different paths depending on the outcome of earlier components.
  • Combine multiple success or failure paths into a single point before continuing.
  • Retry a component automatically if it fails, without manually rerunning the pipeline.
  • Group a sequence of steps into a transaction, and roll back changes if any step fails.
  • Call another pipeline from within the current pipeline, so you can reuse logic across projects.
  • Mark the start and end of a pipeline, including whether it finished successfully or failed.
  • Start or resume warehouse compute resources before running warehouse-intensive steps.