Lineage and organizational context

Publish VegaFlow assets and run evidence into VegaGraph for discovery, impact and governance.

VegaFlow knows how data and AI work runs. VegaGraph connects that operational truth to datasets, code, infrastructure, owners, policies and business outcomes.

What becomes connected

VegaFlow objectVegaGraph representationUseful relationships
Connectionexternal system or data sourcereads_from, writes_to, owned_by
QuickFlow / pipelineworkflow entityproduces, consumes, triggered_by
Published versionversion/build aspectversion_of, built_from, deployed_as
Task or streamoperation entity or field lineagederives_from, writes_to
Compute environmentenvironment/runtime entityruns_on, located_in
Run and attemptexecution evidenceexecuted, failed_at, validated_by

Dataset lineage

When a flow reads and writes known datasets, publish table- and field-level lineage with the immutable flow version as evidence. A later retry should add execution evidence; it should not create a different logical lineage graph unless the mapping changed.

postgres.orders.total
    └─ derives_from → vegadb.raw_orders.net_amount
                          └─ produced_by → orders_ingestion / version 18

Environment context

Keep the logical pipeline canonical and attach deployments, endpoints and runtime facts to a context such as prod/aws/me-central-1. A staging failure then remains visible without making the canonical pipeline look globally unhealthy.

Run evidence

Successful validation, data-quality checks, schema changes and failures can update governed aspects with source run ID, observed time, confidence and evidence link. Secrets, raw payloads and unrestricted logs are never graph attributes.

Impact before change

Before changing a source column or published pipeline version, traverse downstream relationships from the affected field or workflow. The result can include VegaDB tables, dashboards, APIs, services and KPIs—not only the next task in the DAG.

See Contexts and lineage and Governance and change safety for graph rules and impact review.

Vegalake, VegaDB and VegaFlow are trademarks or registered trademarks of Vegalake Inc.

On this page