synapse-step-workflow

To orchestrate complex multi-step actions in the Synapse SDK, use a structured workflow system that manages sequential execution, tracks progress via step weights, and provides automatic rollback on failure.

Automatic rollback ensures data integrity by automatically undoing previously executed steps if a failure occurs during the workflow. This prevents partial updates and leaves the system in a consistent state.

Yes, you can share state and pass data between sequential workflow steps using a shared context object. This allows downstream steps to access outputs and data generated by earlier steps in the execution sequence.

The best way to track progress in a multi-phase execution workflow is by assigning individual weights to each step. The system monitors overall action progress by calculating the completion of these weighted steps.

Yes, the Synapse SDK workflow system supports building data processing pipelines. You can define sequential steps for loading data, transforming it, training a model, and uploading results with automatic rollback if any step fails.

You need a step-based workflow for action development when a complex task requires reliable execution, sequential processing, progress monitoring, and state sharing. It is essential for ensuring data integrity through automatic rollback mechanisms.