Overview

Lineage shows how data flows through your stack—from source systems and warehouse tables, through transformations and jobs, to the dashboards and applications that consume it. Use it to trace quality issues to their root cause, assess the blast radius of a failing job or a planned schema change, and route incidents to the right owner.

Datadog builds lineage automatically from metadata collected through your Quality Monitoring and Jobs Monitoring integrations (Snowflake, BigQuery, Databricks, dbt, Airflow, Fivetran, Looker, Tableau, and others). Anything in the Data Observability Catalog can appear in the graph.

To open Lineage, go to Data Observability > Lineage.

Select anchor assets

Every lineage view centers on an anchor: the single asset whose upstream and downstream neighbors you want to explore. Datadog marks the anchor node with an ANCHOR badge.

To set an anchor, use the search bar at the top of the page:

1. Choose an asset type from the Any asset dropdown (for example, Table, Column, Dashboard, or Job). Leave it set to Any asset to search across all types.
2. Enter the asset name. Datadog searches all connected sources in the Data Observability Catalog.
3. Select a result to anchor the graph.

One anchor

Search for a single asset by name to make it the anchor for the lineage graph.

The graph centers on the selected anchor and shows its upstream dependencies and downstream consumers.

Multiple anchors

Add multiple assets to the search bar to compare related lineage paths in the same view.

Each selected asset is marked with an ANCHOR badge, and the graph shows how their upstream and downstream paths connect.

Search query

Use an attribute query, such as schema:staging, to select a dynamic set of matching assets.

The graph marks every matching asset as an anchor so you can inspect lineage for the full query result set.

The graph renders with the anchors in the center and upstream and downstream neighbors expanding to the left and right.

Navigate the graph

After you set an anchor, the lineage graph renders in the main panel. Upstream dependencies appear to the left; downstream consumers appear to the right. Each node shows the asset’s name, type, source, and basic stats such as row or column count where available.

The toolbar on the right of the canvas provides zoom in, zoom out, Reset view, and Center anchors.

The time selector in the top-right corner (1w, Past 1 Week, and so on) sets the window used to evaluate lineage. Datadog derives relationships from query history and job runs within this window: widen it to surface older or less frequent dependencies, narrow it to show only what’s active.

Lineage Controls

The Lineage Controls panel on the left configures the shape and contents of the graph.

Map, List, and Find

Toggle between Map (the default graph view) and List (a flat, sortable list of every asset in the current slice). Use List to export, copy, or scan a large lineage; use Map to understand structure visually.

The magnifying-glass icon next to the toggle fits the graph to the viewport.

The Find in map search box highlights nodes in the current graph by name. Unlike the top-of-page search, it does not change the anchor—it only locates nodes already on screen.

Depth

Depth controls how many hops of lineage to load on either side of the anchor.

• The left selector sets upstream depth (levels of parents).
• The right selector sets downstream depth (levels of children).
• Set either to ∞ to load all available hops in that direction.

Increase depth to find a distant root cause or downstream consumer. Decrease depth when the graph is too dense to navigate.

Filter

The Filter section controls which asset types are displayed. For Snowflake, the available types are Column and Table; BI integrations add dashboards and reports; jobs add tasks and DAGs. The number next to each type shows how many of those assets exist in the current slice.

Filter when the slice contains the right assets but the graph is too noisy. For example, when scoping the blast radius of a column change, uncheck Table to remove table-level clutter and leave Column checked.

Filtering does not change the anchor or the depth—it only hides nodes from the rendered graph.

Group by

Group by sets the level of aggregation. Available levels depend on the source. For Snowflake, you can group by Accounts, Databases, Schemas, Tables, or Columns.

Grouping is most useful for zooming out: group by Schemas to see how data flows across a warehouse, then drill down to Tables or Columns after you find the area of interest.

Common workflows

Root cause analysis

When a downstream asset—a dashboard, a model, an ML feature—is broken or stale, lineage helps you walk backward to the source.

1. Anchor on the broken asset (a Looker dashboard, a Snowflake table, a dbt model).
2. Set downstream depth to 0 to focus on upstream assets.
3. Group by Tables for the broad structure; switch to Columns if the issue is at column level.
4. Step backward through upstream nodes. Failures, freshness anomalies, and schema changes flagged by Quality Monitoring or Jobs Monitoring appear as status indicators on the graph.
5. Open a flagged node to jump to its quality monitors, recent job runs, or schema history.

Impact analysis (blast radius)

Before changing or dropping a column, table, or model, use lineage to see what depends on it.

1. Anchor on the asset you plan to change.
2. Set downstream depth to ∞ and upstream depth to 0.
3. Filter to the asset types you care about—for example, leave dashboards and reports visible to identify affected BI consumers.
4. Switch to List view to export the full list of affected assets or share it with the owning teams.

Tracing a column end-to-end

Most integrated sources support column-level lineage.

1. In the search bar, change the asset type to Column and search for the column to trace.
2. Anchor on the result.
3. Group by Columns to keep the graph at column granularity.

Further Reading

Additional helpful documentation, links, and articles:

Data CatalogDOCUMENTATION Quality MonitoringDOCUMENTATION Jobs MonitoringDOCUMENTATION Understanding data lineageBLOG