Skip to main content

Cross-workspace extraction issues

This guide helps you troubleshoot failures in cross-workspace extraction for Databricks. It covers errors related to service principal changes, enrichment migration, lineage, and new workspace inclusion.

Scheduled run fails during preflight check​

Error

Preflight check failed: missing required permissions for catalogs

Cause​

The catalogs from the revoked workspace remain in the include or exclude filters. Since the service principal no longer has permissions, the workflow can't proceed.

Solution​

  1. Remove catalogs from the revoked workspace in the include or exclude filters.
  2. Update the connection before the next scheduled run.
  3. Retry the workflow.

Workflow fails due to high percentage of deleted assets​

Error

Workflow failed: deleted assets exceeded 80% circuit breaker threshold

Cause​

If the revoked workspace contained a large portion of metadata, its sudden removal triggers a mass deletion of assets. When the percentage of deleted assets is greater than 80%, the workflow fails by design (circuit breaker behavior).

Solution​

  1. Review the percentage of deleted assets in the workflow logs.
  2. Manually increase the circuit breaker value for the workflow.
  3. Rerun the workflow to complete the extraction.

Enrichments missing after migrating to cross-workspace connection​

Error

Descriptions, tags, owners, or glossary terms not visible after migration

Cause​

When migrating from multiple workspace-specific connections to a single cross-workspace connection, enrichments aren't automatically transferred unless qualified names are preserved.

Solution​

  1. Deploy the new cross-workspace connection and run an initial crawl.
  2. Run the asset-export workflow on both the old and new connections to generate metadata extracts.
  3. Match assets by qualified names and transfer enrichments (descriptions, tags, owners, glossary terms) using a spreadsheet.
  4. Run asset-import to ingest the merged metadata into the new connection.

Lineage missing in cross-workspace extraction​

Error

No lineage generated across workspaces

Cause​

Cross-workspace extraction only collects metadata. Lineage requires a Databricks miner connection to process and connect query activity to the extracted metadata.

Solution​

  1. Configure a Databricks miner connection for the cross-workspace setup.
  2. Run the miner workflow to build lineage across extracted assets.
  3. Verify lineage in Atlan after the miner job completes.

Newly added workspace not included in extraction​

Error

Workspace not listed or metadata not extracted from new workspace

Cause​

The service principal doesn't have permissions on the newly added workspace, preventing the crawler from accessing metadata.

Solution​

  1. Assign the existing service principal the required permissions on the new workspace.
  2. Re-run the cross-workspace crawler to include the new workspace.
  3. Verify that catalogs from the workspace appear in the extraction results.

Need help​

If you need assistance after trying the steps, contact Atlan support: Submit a request.