Lineage
Cross-platform lineage — one graph, many tools
Connect dbt, Airflow, and warehouse tables so blast radius is not trapped in a single UI. Includes a concrete lineage example.
Lineage example (orders → revenue)
A practical lineage example mid-market teams hit every week:
Airflow: orders_ingest → raw.orders dbt: stg_orders → fct_orders → mart.daily_revenue → Looker: exec_revenue_dashboard Consumer contract: max staleness 6h on mart.daily_revenue
When raw.orders lags, you need the whole chain — not just the dbt node or the Airflow task in isolation. That is cross-platform lineage.
What the catalog stores
Dataset nodes
Tables and models with types and connection refs.
Pipeline edges
From → to with the job that produced the link.
Owners & schedules
Who to page; when it should have landed.
Check results
Freshness/volume tied to the same graph.
Passport status
Which share links are still trustworthy.
CI impact
Downstream + contracts at risk on a PR.
Build the graph
- 01
Import tools you already use
dbt manifest and/or Airflow DAG.
- 02
Open lineage explorer
Click a mart; walk upstream and downstream.
- 03
Gate PRs on impact
Comment blast radius before merge.
Cross-platform lineage FAQ
Examples and how the graph stays honest.
What is cross-platform lineage?+
Cross-platform lineage connects datasets and jobs across tools — e.g. a Postgres source loaded by Airflow, transformed in dbt, and consumed by a BI dashboard — so impact analysis is not trapped inside one product's UI.
Can you show a lineage example?+
A common example: raw.orders (ingestion DAG) → stg_orders (dbt) → mart.daily_revenue (dbt) → executive dashboard. DataXPipe stores those edges in the catalog, attaches freshness checks to mart.daily_revenue, and shows blast radius when the source lags.
How does DataXPipe build cross-platform lineage?+
Import dbt manifests and Airflow DAGs, or declare edges in YAML pipeline specs. Query lineage per dataset via API or the lineage explorer; use impact analysis in CI for PR blast radius.
See blast radius before merge
Import dbt or Airflow, open the lineage explorer, and run PR impact analysis on your next change.