- One query root per environment:
environment(slug: "production")exposes the project two ways, thedefinitionstate as declared in code and theappliedstate joined with the latest run results. - Full resource coverage: models, sources, tests, seeds, snapshots, exposures, metrics, macros, groups, and semantic models, each with filters, cursor pagination, and per-node
parents,children, andancestorslineage. - Execution and quality metadata: last run status and timing, per-model run history, test failures with failing row counts, source freshness checks against declared SLAs, and warehouse catalog stats.
- Whole-DAG lineage in one call: the
lineagefeed returns every node with its parent edges, and exposures carry a computed health rollup across all of their upstream dependencies. - dbt Cloud compatible query shapes: if you have tooling built against the dbt Cloud Discovery API, the query patterns carry over with minimal changes.