Skip to main content
Workflows let you orchestrate multi-step data processing pipelines in CDF. A workflow defines a sequence of tasks that are executed in order, with support for conditional logic and error handling. You create workflows as named objects, then add versions that define the actual task graph.

​
Overview

Workflows are the top-level container for your automation logic. A workflow has a name and metadata, but the execution logic lives in workflow versions. This separation lets you iterate on workflow logic while keeping a stable workflow identity.

​
Workflow structure

Each workflow contains:
  • Metadata – Name, description, and optional labels
  • Versions – One or more workflow versions that define the task graph
  • Configuration – Settings that apply across versions

​
Creating workflows

You create a workflow by providing a name and optional description. After creation, you add workflow versions to define the tasks and their execution order. A workflow can have multiple versions; you typically run a specific version when triggering an execution.
Workflow versions are immutable. To change the task graph, create a new version rather than editing an existing one.

​
Use cases

Workflows are useful for:
  • ETL pipelines that transform and load data
  • Scheduled data synchronization jobs
  • Event-driven processing triggered by CDF or external events
  • Multi-step validations and approvals
Last modified on April 23, 2026