Table of Contents

Data Pipeline Runs

Learn how to view, filter, analyze, and cancel data pipeline runs.

Overview

The Data Pipeline runs page lists pipeline runs with their status and outcome. From it you can refresh individual runs, open a detailed analysis, and cancel runs that are still in progress.

In the sidebar, select Data > Data Pipeline Runs. The list is sorted by start time, newest first, and loads 20 runs at a time as you scroll ("Loading more pipeline runs...").

Runs Table

Column Description
Data Pipeline The pipeline the run belongs to.
Activation An icon for the trigger type: person for Manual, clock for Schedule, bolt for Event. Hover to see the type.
Start Time When the run started. Sortable.
Duration Run duration as mm:ss, or In progress.
Status Completed or In progress. Sortable.
Success For completed runs: Success (green), Success with warnings (amber), or Failure (red). In-progress runs show a dash. Sortable.
Completed Stages Stages that have finished.
Failed Stages Stages that failed.
Actions Refresh run, Analyze run, and Cancel run.

Select the refresh icon at the top right of the page to reload the whole list.

Filter Runs

The filter panel above the table has these filters:

Filter Options
Pipeline Any, or a specific pipeline.
Activation Any, Manual, Schedule, or Event. The page opens with Manual selected.
Status Any, Completed, or In progress.
Success Any, Success, or Failure.
Period Any, Last 1 hour, Last 24 hours, Last 7 days, Last 30 days, or Custom range.

Select More filters to show:

  • User: Search by UPN and select a user to show only runs that user triggered.
  • Custom range: From and To date and time pickers. Shown when Period is Custom range.

Active filters appear as chips below the panel (for example, "Status: Completed"). Select the x on a chip to remove that filter, or select Clear all to reset every filter to Any.

Run Actions

Refresh a Run

Select the Refresh run icon to reload a single run's status. If the run no longer matches the current filters, it is removed from the list.

Analyze a Run

Select the Analyze run (chart) icon to open the Data Pipeline Run Analysis dialog. It contains the run header (Run ID, Owned By, Triggered By, Start Time, End Time, Total Duration) and these sections:

Section Contents
Errors Error messages recorded for the run. Shown only when there are errors.
Execution Timeline Timeline bars for the pipeline run, each stage, and each stage's work items, with a legend: Pipeline Run, Stage (all successful), Stage (has failures), Has Warnings, and Work Item Extent.
Invocation Parameters Canonical Run ID, Processor, Trigger, Triggered By, and each trigger parameter value. Each value has a copy icon.
Stage Metrics Analysis Per-stage work item counts (completed, successful, warnings), duration, and a Work Items table when telemetry is available.

Select Close to close the dialog.

Cancel a Run

The Cancel run icon is enabled only for runs that are in progress.

  1. Select the Cancel run icon.
  2. In the Cancel Pipeline Run dialog, select Cancel Run. Select Keep Running to leave the run alone.

Processing stops and the run is marked as cancelled. A notification confirms the cancellation.

Common Scenarios

Find failed runs: Set Activation to Any, set Success to Failure, and open Analyze run on a run to see its Errors and Failed Stages.

Watch active runs: Set Status to In progress and use Refresh run on the runs you care about.

Review recent activity: Set Period to Last 24 hours and Activation to Any.

Find runs you started: Under More filters, search for your UPN in User.

Troubleshooting

  • A run you expected isn't listed: The page opens filtered to Manual runs. Set Activation to Any, and check the other filter chips. Viewing runs requires dataPipelines/write on the pipeline. With Pipeline set to Any, that permission is required at the instance scope, so a Reader role alone does not list runs. See Data Pipeline Permissions.
  • User search shows "Access Denied": You don't have permission to search security principals.
  • No stage metrics or timeline: The run may not have started any stages yet, or telemetry isn't available for it.