> ## Documentation Index
> Fetch the complete documentation index at: https://docs.vegalake.com/llms.txt
> Use this file to discover all available pages before exploring further.

# Lineage and impact

> Traverse upstream and downstream entity or field relationships and assess impact.

# Lineage and impact

Lineage is a constrained traversal of typed relationships. Use it to understand where data or behavior comes from, where it goes, and what may be affected by a change.

## Choose the right granularity

* **Entity lineage** follows relationships among resources such as tables, models, services, or jobs.
* **Field lineage** follows field-reference edges, such as column-to-column derivation.

Field lineage can be absent even when entity lineage exists. Confirm that field references and their edges were ingested.

## Direction and depth

* **Upstream** finds inputs, dependencies, or producers.
* **Downstream** finds consumers, dependents, or outputs.
* **Depth** limits how many edge hops are traversed.
* **Relationship filters** constrain traversal to meaningful edge types.

Start with a shallow depth and a specific relationship set. Expand only when the result remains interpretable.

## Impact views

Specialized endpoints can project KPI impact or type-impact previews from the same graph. Treat impact as evidence from current edges and contexts, not proof that every external dependency was ingested.

## Graph constraints

Relationship types can enforce endpoint types, cardinality, and acyclic behavior. Those constraints improve lineage quality but do not automatically discover dependencies. Ingestion is outside VegaGraph, and agent execution is outside VegaGraph.

## Troubleshooting

Check organization/workspace, selected context, search result identity, traversal direction, depth, relationship filters, field/entity granularity, permissions, and the existence of edges. Preserve the lineage search state when sharing a reproducible investigation, but do not include credentials or secret values.