Connector Contracts
Early pilots use customer-contained Databricks system tables and file exports before authenticated vendor API sync is enabled.
Supported Contracts
- Databricks system tables through
DATABRICKS_WAREHOUSE_ID. - BI dashboard metadata through
BRICKCOST_BI_METADATA_PATH. - Workflow metadata through
BRICKCOST_WORKFLOW_METADATA_PATH. - dbt manifest through
BRICKCOST_DBT_MANIFEST_PATH. - Airflow or Dagster metadata through
BRICKCOST_ORCHESTRATION_METADATA_PATH. - Git metadata through
BRICKCOST_GIT_METADATA_PATH. - Ticket metadata through
BRICKCOST_TICKET_METADATA_PATH.
Readiness API
Enterprise deployments expose GET /api/connectors so the diagnostic team can
see which optional connectors are configured and which records were parsed.
BI Metadata Import
The Databricks App includes a Connector Setup page for pasting a dashboard metadata JSON export. Once imported, customer BI metadata replaces the bundled synthetic sample evidence in diagnostics, reports, and Genie readiness findings.
The API equivalent is POST /api/bi-metadata/import. The import stores metadata in
the customer's Brickcost storage schema and refreshes the query-intent and transformation
diagnostic snapshots in the background without sending metadata to an external service.
Customers can validate import shape first with POST /api/bi-metadata/validate.
Validation returns row-level warnings for missing owners, tables, metrics, duplicate ids,
and unsupported risk values.
Optional dbt, workflow, ticket, orchestration, and Git file exports are incorporated into audit evidence, operating-graph nodes, and MCP connector context so pilots can trace owners, schedules, incidents, and remediation repositories without moving metadata outside the customer's workspace.