Every asset in the Data Catalog carries a Lineage tab that traces where its data comes from and where it flows next. Instead of guessing from naming conventions or asking around, you can open the graph and see the actual chain of tables, fields, and pipelines that produce and consume an asset — and use that chain to plan changes safely or track down bad data at its source.

Overview

The Lineage tab renders an interactive flow graph centered on the asset you opened. It's laid out left to right: upstream sources — where the asset's data originates — sit on the left, and downstream consumers — what's built from it — sit on the right. Nodes in the graph represent data assets, and the lines connecting them represent the data flows between those assets. Lineage is available on both tables and individual fields, so you can trace an entire table or narrow the view to a single column.

You'll find it at Data Catalog › open an asset › Lineage.

Exploring the graph

The graph starts small and grows as you explore it. Expand a node's upstream or downstream side to load the next hop in either direction, and keep expanding to walk the chain as far as you need. Clicking any node opens a details drawer with that asset's overview, so you can inspect assets along the chain without losing your place in the graph.

For table nodes, expand the node's fields to see its columns listed inside it. Selecting a field highlights field-level lineage — the specific column-to-column flows feeding that field, rather than just a table-to-table connection. This is the difference between knowing that two tables are related and knowing exactly which column produced which.

Where the edges come from

Lineage edges aren't manually drawn — they're recorded from the pipelines and integrations that actually move data between assets. When data flows through a pipeline, the mapping from source to destination, down to the field level, is captured against the catalog automatically. Any edge backed by a pipeline is clickable: select it and choose "Click to see pipeline details" to jump straight to the pipeline that created the flow. Lineage can also be registered programmatically, so a custom automation can record a flow using the built-in Data Catalog operations even when a standard pipeline isn't involved.

Using lineage for impact analysis and root-cause tracing

Lineage exists to support two everyday jobs:

• Impact analysis — before changing or dropping a table or column, walk downstream from it to see every asset that depends on it and would break as a result.

• Root-cause tracing — when numbers look wrong in a report or dashboard, walk upstream from it to find the source asset feeding the bad values.

Field-level highlighting makes both tasks precise. Rather than guessing which downstream table is affected based on its name, you can select the exact field you're changing and see exactly which columns in which assets consume it.

Notes

• Lineage only appears where a recorded flow touches the asset. If an asset was synced from a source but no pipeline reads from or writes to it — or its pipelines ran before lineage capture was enabled — the graph has nothing to draw.

• An empty graph fills in over time, either as pipelines involving the asset run going forward or as lineage is registered explicitly through automation.

• Edges without an associated pipeline (for example, those registered via automation) won't have a "Click to see pipeline details" option, since there's no pipeline record behind them.