Recurring runs

A recurring run executes a pipeline on a repeating schedule or trigger. Use recurring runs for workloads that need to run regularly (such as nightly retraining, daily data processing, or periodic model evaluation) without manual intervention each time.

Create a recurring run

1. Select Pipelines in the left sidebar, then select the Recurring runs tab.

2. Select + Create recurring run.

3. In the Create a recurring run dialog:

   • Pipeline: select the pipeline to run.
   • Pipeline version: select the version of the pipeline.
   • Run name: enter a name that identifies this schedule.
   • Description: optional notes about this recurring run.
   • Experiment: select the experiment to associate runs with.
   • Run parameters: fill in any parameters defined by the pipeline.
   • Trigger: choose how the run repeats:
     • Periodic: runs every fixed interval (minutes, hours, or days).
     • Cron: runs on a cron expression schedule.
   • Start date / End date: optionally bound when the schedule is active.
   • Maximum concurrent runs: limits how many runs can execute at the same time if a previous run is still in progress.

4. Select Create.

The recurring run appears in the Recurring runs list with its next scheduled execution time.

Enable or disable a recurring run

You can pause a recurring run without deleting it. Select the overflow menu (⋮) next to the recurring run and select Disable. Select Enable to resume it.

Delete a recurring run

Deleting a recurring run cancels all future scheduled executions. Runs that have already completed are not affected and remain visible in the Runs tab.

To delete a recurring run, select the overflow menu (⋮) next to it and select Delete.

Next steps

• Monitor individual runs triggered by this schedule.
• Browse artifacts produced by recurring runs.