OPENLINEAGE-SPEC
Cross-platform data lineage standard.
Definition
OpenLineage defines an event model with runEvents (START/COMPLETE/FAIL), jobs (name, namespace), inputs/outputs (datasets), facets (structured extensions). Native integrations: Airflow, Spark, dbt-ol, Dagster, Flink, Trino. Consumed by Marquez, Atlan, DataHub, OpenMetadata.
Origin
Project created by WeWork in 2020 ; donated to LF AI Foundation in 2021 ; v1.0 released June 2022.
Example in context
A data team sends Airflow + dbt runs to OpenLineage then visualises lineage in Marquez.
Related terms
- dbt — native integration.