Executions

Browse pipeline step execution records in BullSequana AI.

Agentic Friendly

Executions are records of individual pipeline step runs. Each time a step in a pipeline runs, an execution record is created that captures what inputs the step received, what outputs it produced, and its final state. Use the Executions view to audit the lineage of any artifact or to investigate a failed step.

Browse executions

Select Pipelines in the left sidebar, then select the Executions tab.

The Executions list shows every step execution across all runs, along with:

  • Name: the step name as defined in the pipeline.
  • Type: the execution type (for example, system.ContainerExecution).
  • State: the final state of the step (COMPLETE, FAILED, or CACHED).
  • Created: when the execution was recorded.

View execution details

Select an execution to open its detail page. The detail page shows:

  • Properties: metadata about the execution, including the run that triggered it.
  • Input artifacts: artifacts consumed by this step.
  • Output artifacts: artifacts produced by this step.

A CACHED state means the platform reused the output from a previous identical execution rather than re-running the step. Caching reduces execution time when upstream inputs have not changed.

Relationship to runs

Executions map one-to-one with pipeline steps within a run. To see all steps from a single run together, open the run detail page from the Runs tab.

Next steps

On this page