Pipeline Run Details

Click a pipeline name on the Pipelines listing page to open the Pipeline Run Details page. This page gives you a detailed view of historical runs, the pipeline structure, and the events and automations associated with a selected run.

The page has three main areas:

• Summary Panel
• Lineage Graph
• Details Panel

Summary panel

The header at the top of the page shows the key information for the selected run, including:

• Pipeline name
• Run timestamp
• Execution time, along with a comparison to the historical baseline
• Data reliability policies ratio, showing the number of checks passed compared to the total number of checks evaluated during the run

Click the run timestamp to open the Pipeline Run History dropdown. From there, you can select and compare two runs.

Pipeline tags

The tag bar at the top of the page shows the pipeline's tags. This includes tags that you add in ADOC and Airflow DAG tags that ADOC ingests from OpenLineage. Each tag is shown as key: value.

To add tags to a pipeline:

1. Click the tag icon in the tag bar.
2. In the Add Tags dialog, enter a Key Name and a Value.
3. To add another tag, click Add Row.
4. Click Update.

Lineage graph

The center of the page contains the interactive lineage graph. This graph shows the jobs and data assets involved in the selected pipeline run.

The graph includes:

• Job nodes, shown as white icons.
• Asset nodes: Blue nodes are identified assets. They match an asset in the ADOC catalog and link to its profile, quality history, and observability data. Gray nodes are unidentified assets. A pipeline reported reading or writing them, but they aren't in the ADOC catalog, so they don't link to a catalog entry. To identify an asset, add its data source to ADOC and crawl it. The asset then appears in blue.

You can click, pan, search by name, and zoom to explore the graph. Controls in the upper-right corner let you:

Filtering the graph

Click the eye icon in the upper-right corner of the graph to open a panel with the following options:

• Only Job Nodes: Hide data assets and show only job nodes.
• Only Errored Nodes: Show only nodes with an error state, dimming everything else in the graph.
• Related Pipelines: Show other pipelines that read from or write to the same underlying asset as the current pipeline. It is turned off by default.
• Linked Pipelines: Show pipelines connected to the current pipeline through cross-layer lineage stitching. It is enabled by default.
• Child Nodes: Expand asset nodes inline to show their column- or field-level schema.

After selecting your options, click Apply. This preference applies across all pipelines you view, not just the current one.

A Nodes by type legend in the lower-right corner of the graph shows a count of each node type present (for example, Assets and Jobs).

Exploring linked pipelines

When Linked Pipelines is enabled, a pipeline triggered by a task in the current pipeline — for example, a Databricks Spark job, Trino query, AWS Glue job, or dbt Cloud transformation started by an Airflow task — appears as an expandable node in the graph.

• Click the expand icon on a linked pipeline node to reveal that pipeline's own upstream and downstream data assets inline, without leaving the current view.
• Click any asset node to navigate to that asset's detail page in the ADOC catalog.
• Click Collapse to return to the top-level view.

This lets you trace a complete path — from the orchestrating pipeline, through the processing job it triggers, to the catalog-registered assets that job reads from and writes to — in a single graph.

Tracing dependencies

Use highlighting to follow dependencies through the lineage graph without losing your place.

• Hover: Hover over a node to temporarily highlight its connected edges. The highlighting disappears when you move the pointer away.
• Highlight nodes: Command+click (macOS) or Ctrl+click (Windows) one or more nodes to keep them highlighted. ADOC also highlights their directly connected nodes and the edges between them, and dims the rest of the graph. The highlighting stays while you pan or zoom, and when you click a highlighted node to open its tab in the Details panel.
• Isolate selected nodes: To focus on the highlighted nodes alone, hold S. ADOC dims their connected nodes and hides the edges until you release the key. To keep the selected nodes isolated while you move around the graph, click Isolate the selected nodes in the upper-right controls instead.

To clear highlighting and isolation, click an empty area of the graph.

Viewing node details

Click any job or asset node in the graph to open a tab for that node in the Details panel below the graph, alongside the Timeline and All Automations tabs. You can open multiple node tabs at once and close each individually.

To keep a node and its connections highlighted while you explore the graph, see Tracing dependencies.

A node's tab shows different information depending on its type:

• Job nodes show Data Reliability and Metadata tabs, including any Executed Data Reliability Policy and Executed Profiles linked to that job.
• Asset nodes show Data Reliability, Metadata, and Sample Data tabs, including Executed Data Reliability Policies, Executed Profiles, and Related Data Reliability Policies linked to that asset.

Hard linked and soft linked policies

A policy can exist in a pipeline and execute itself in two ways:

1. Hard Linked: This policy type exists within the pipeline. Policy is executed automatically with each pipeline run, just like any other job. Here the score will be for the data set just processed during that run, leading to accurate results.
2. Soft Linked: This policy type exists independent of this pipeline in association to the asset. It have it's own schedule which can differ from the Pipeline run schedule. Scores are populated from the latest execution of policy after any pipeline run. Score will be generated based on the policy's data incremental strategy. Which will lead to indicative scores for the data processed during any pipeline run.

Details panel

The panel below the lineage graph contains multiple tabs with detailed information about the run, including a tab for each job or asset node you select in the lineage graph (see Viewing node details above)."

Timeline

The Timeline tab displays the span timeline chart, which shows the duration and sequence of each job or span in the pipeline run.

Each horizontal bar represents a span, and the length of the bar reflects its duration. Use this view to identify bottlenecks or long-running tasks within the run.

Event list

Click View all events from the Timeline tab to open the Event List.

The Event List displays a time-stamped record of all events that occurred during the run, including:

• Pipeline start and completion events
• Task start and completion events
• Task failure events, including error details
• Input and output asset events

Automation

The Automation tab lists all data reliability automations executed during the run. The Executed Policies section shows any profiling, data quality, or reconciliation automations that were triggered by the pipeline outcome.

Creating an Automation

To create a new automation, navigate to the pipeline's Automation tab from the pipeline details page and click Create Automation. The When should this data quality check run? dialog guides you through the configuration:

1. Execute automation on completion of:  Select the trigger event, such as Pipeline, and the specific completion condition that should trigger the automation.
2. Automation type:  Choose one of the following:
   • Profiling: Discover and profile a catalog asset after the trigger completes.
   • Data quality:  Evaluate a data quality policy when the trigger completes.
   • Reconciliation: Run a reconciliation or equality policy when the trigger completes.
3. Select an asset: Choose the asset the automation should run against. This field is a searchable dropdown and is automatically scoped to the assets associated with the selected pipeline, so you do not need to know or type the exact asset name. To search across all assets in the catalog instead of only those linked to the pipeline, turn on Search all assets.

4. Delay in Minutes: Optionally set a delay, in whole minutes, before the automation runs after its trigger completes. This applies to all automation types (Profiling, Data Quality, and Reconciliation) and is useful for staggering automations that would otherwise start at the same time on a shared Data Plane cluster, reducing resource contention. Leave the field at 0 to run the automation immediately when the trigger completes.
5. How should each run execute?: Choose the execution strategy:
   • Full: Run on the full dataset every time.
   • Incremental: Run on new or changed data, based on the incremental strategy configured on the asset or policy.

Click Save Automation to complete the configuration.

The configured delay is visible in the Executed Policies section for each automation run, in the Delay (min) column, alongside the Scheduled Start time showing when the delayed run is set to fire.

What’s next

To configure alerts for duration or failure thresholds on this pipeline, see Manage Pipelines. To explore lineage across systems and assets, see the Lineage documentation.