# Runs (/docs/guides/developer-workspace/pipelines/runs)



A run is a single execution of a pipeline with a specific set of parameter
values. Each run is immutable: its inputs, outputs, and logs are preserved
exactly as they were when the run completed. You can review a run at any
time to reproduce or compare results.

Start a run [#start-a-run]

1. Select **Pipelines** in the left sidebar, then select the **Runs** tab.

2. Select <kbd>+ Start run</kbd>.

3. In the **Start a run** dialog:

   * **Pipeline**: select the pipeline to run.
   * **Pipeline version**: select the version of the pipeline. The latest
     version is selected by default.
   * **Run name**: enter a descriptive name for this run.
   * **Description**: optional notes about this run.
   * **Experiment**: select the experiment to associate this run with.
   * **Run parameters**: fill in any parameters defined by the pipeline.

4. Select <kbd>Start</kbd>.

The run appears in the **Runs** list with a **Running** status.

Monitor a run [#monitor-a-run]

Select a run in the list to open its detail page. The detail page shows:

* The **graph**: each step highlighted with its current status (pending,
  running, succeeded, or failed).
* The **Run output** tab: artifacts and metrics produced by the run.
* The **Config** tab: the parameter values used for this run.

Select any step in the graph to see its logs and output in the side panel.

Clone a run [#clone-a-run]

Cloning a run creates a new run pre-filled with the same pipeline version
and parameter values. Use this to repeat a run with minor parameter
adjustments.

1. Open the run detail page.
2. Select <kbd>Clone run</kbd>.
3. Adjust any parameters, then select <kbd>Start</kbd>.

Next steps [#next-steps]

* [Set up a recurring run](recurring-runs) to execute the pipeline on a
  schedule.
* [Browse artifacts](artifacts) produced by your runs.