Data Lineage¶
Go to Data Observability > Data Lineage to view data sources and downstream dependencies around data assets, jobs, or fields. You can start from a target object and continue investigating along its upstream and downstream relationships.
Selecting Anchors¶
An anchor is the starting point of the lineage analysis. It can be a single asset or multiple assets.
When you first enter Data Lineage, the page shows a relationship diagram. After selecting anchors using Search assets or filter by tag at the top, the page displays the corresponding lineage relationships.
- Search or filter assets in the search box at the top. For search syntax, see Search.
- Click the asset you want to view to add it as an anchor.
- Adjust the direction, levels, and view based on your investigation goals.
You can continue adding other assets. After you remove all anchors, the page returns to the default diagram page.
The search dropdown shows the number of matches and the list of assets. You can add them one by one, or click Add All to set all matching assets as anchors.
When the page indicates that the anchor limit has been exceeded, add more filter conditions or reduce the number of selected anchors.
Before viewing field-level lineage, make sure field lineage collection is enabled.
Adjusting the Lineage View¶
| Control | Description |
|---|---|
| Direction | Upstream and Downstream shows both sides at the same time, Upstream traces origins, and Downstream shows dependent objects. Default: Upstream and Downstream. |
| Levels | You can select 1, 2, 3 levels, or All. Default: 2 levels. In upstream/downstream modes, the graph expands to the corresponding sides. |
| Grouping | Group by platform, database, or schema, or do not group. Assets without a grouping attribute are shown as "Not Reported". |
| View | Switch between map and list, preserving anchors, direction, and levels. |
How Levels Are Calculated¶
Each actual connection counts as one level. For example:
Starting from the input data asset, reaching the job is 1 level, and reaching the output data asset is 2 levels.
After selecting "All", the page continues to expand relationships. If you see a prompt about limited results or boundary nodes, reduce the levels or the number of anchors and view the results in segments.
Map and List¶
Map¶
The map shows upstream objects, anchors, and downstream objects. Anchors have a dedicated marker, and nodes display name, type, platform, and availability status. Connecting lines indicate collected dependency directions.
You can zoom, reset the scale, view in full screen, and select objects for further analysis. When investigating around another object, you can set that object as a new anchor.
List¶
The list displays object, type, platform, direction, and recent relationship information, making it suitable for item-by-item verification. The map and the list display the same set of relationships.
Viewing Relationship Details¶
Select a connecting line to view relationship details, including relationship type, most recent record time, and associated Run ID. To view the corresponding run, you can search by Run ID in the job details.
Common relationships include:
| Relationship | Direction and Meaning |
|---|---|
| Read relationship | Data asset → Job. The job reads the asset. |
| Produce relationship | Job → Data asset. The job produces the asset. |
| Field source relationship | Upstream field → Downstream field. Shows the source of the field. |
Accessing Lineage from Details¶
| Entry | Initial Scope |
|---|---|
| View Full Lineage in data asset details | Current asset, upstream and downstream, 2 levels. |
| View Lineage in job details | Current job, upstream and downstream, 2 levels. |
| View Lineage for This Run for inputs and outputs | Direct inputs and outputs of the selected run, with fixed direction and levels. |
| View Downstream Impact in run details | Current job, downstream, 2 levels. |
| View Downstream Impact for runs associated with a data asset | Current asset, downstream, 2 levels. |
Tracing Origins and Analyzing Impact¶
Trace Origins: Select the target data asset as the anchor and choose Upstream, then examine run records along the input data and producing jobs.
Analyze Downstream Impact: Select a failed job or a data asset to be inspected as the anchor and choose Downstream, then view dependent objects. Enter their details to check run status and error information, and confirm the impact.