AWS Glue Data Catalog

AWS Glue Data Catalog data catalog

Glue Data Catalog stores table schemas and partitions for AWS analytics services, and federates Databricks and Snowflake catalogs through Iceberg REST.

Compare: data catalog tools

• Automated Metadata Harvesting: Crawlers scan Amazon S3, JDBC databases and open table formats (Iceberg v3, Delta 4.0, Hudi as of 2026-09-25) and register schemas and partitions automatically (docs).

• Search & Discovery: Tables, views and columns are browsable and searchable in the console and through the SearchTables and new SearchAssets APIs (docs).

• AI-Powered Search: Semantic search (Preview, 2026-06-17) finds tables by business meaning through the Glue Search API, in four regions (docs).

• Custom Fields & Templates: Custom metadata fields (forms) standardize key-value metadata attached to assets (Preview, 2026-06-17) (docs).

• Business Glossary: Business glossaries with terms that associate to tables went into Preview on 2026-06-17, without documented steward approval workflows (docs).

• Metadata Propagation: Lake Formation LF-Tags assigned to a database are inherited by its tables and columns, with per-resource overrides (docs).

• Usage & Popularity Statistics: No per-table query counts or popularity rankings are documented in the Data Catalog (docs).

• Asset Certification: No certification or deprecation status for catalog assets is documented (docs).

• Catalog Usage Analytics: No catalog adoption or coverage dashboard is documented; metadata can be exported to S3 Tables (Preview) for DIY analysis (docs).

AWS Glue Data Catalog data stewardship

• Metadata Automation: Catalog changes emit EventBridge events that customers can wire to automation, but no built-in metadata rules are documented (docs).

• Critical Data Elements: No critical data element registry is documented (docs).

• Review & Approval Workflows: No approval workflows in the Glue Data Catalog; subscription approvals live in Amazon SageMaker Catalog (docs).

• Governance Campaigns: No bulk steward campaigns are documented (docs).

• Governance Scorecards: No governance maturity scoring is documented (docs).

AWS Glue Data Catalog data classification

• Data Policy: Lake Formation enforces database, table, column, row and cell-level permissions on Data Catalog resources, including LF-Tag-based access control (docs).

• Data Masking: Lake Formation hides columns and filters rows and cells, but value-level masking isn't documented (docs).

• Sensitive Data Discovery: Catalog tables aren't classified automatically; sensitive-data detection is an AWS Glue ETL transform, and Amazon Macie is a separate service (docs).

• Native Policy Enforcement: Lake Formation permissions are enforced by AWS engines (Athena, EMR, Redshift Spectrum, Glue); catalog federation brings remote Iceberg catalogs into AWS, with no documented policy pushdown into Snowflake or Databricks (docs).

AWS Glue Data Catalog data compliance

• Audit Trail: AWS CloudTrail records Glue Data Catalog API calls and Lake Formation data access events (docs).

• Compliance Reporting: No compliance reports or evidence packs are documented (docs).

• Regulatory Framework Mapping: No mapping of assets to regulations is documented (docs).

• Data Activity Monitoring: CloudTrail logs access events, but the Data Catalog documents no access-pattern analysis (docs).

• ROPA: No GDPR Article 30 register (docs).

• Data Retention Policies: No data retention or deletion policies in the Data Catalog; Iceberg snapshot cleanup is table maintenance, not compliance retention (docs).

• Data Subject Rights: No DSAR workflow (docs).

AWS Glue Data Catalog data marketplace

• Data Mesh Support: Lake Formation cross-account sharing is a common data mesh building block, but the Data Catalog has no domain or federated-governance model (docs).

• Self-Service Access Requests: No access request flow in the Glue Data Catalog; subscription requests live in Amazon SageMaker Catalog (docs).

• Data Products: No data products in the Glue Data Catalog; they're an Amazon SageMaker Catalog feature (docs).

• Subscription Management: No subscriptions in the Glue Data Catalog (docs).

AWS Glue Data Catalog data quality

• SQL Table Tests: AWS Glue Data Quality runs DQDL rules, including custom SQL, on Data Catalog tables and Athena views (billed separately) (docs).

• Data Health Scores: AWS Glue Data Quality reports a data quality score per ruleset run on a catalog table (docs).

• Anomaly Detection: AWS Glue Data Quality detects anomalies in data statistics using ML (docs).

• Data Profiling: Column statistics (nulls, distinct values, min/max, lengths) are computed and stored in the Data Catalog, billed per DPU-hour (docs).

• Data Contracts: No producer-consumer contract feature is documented (docs).

AWS Glue Data Catalog AI governance

• AI Agent Access Governance: Agents read the catalog under IAM and Lake Formation permissions; delivery of policy context alongside agent responses isn't documented (docs).

• AI Asset Inventory: The Glue Data Catalog stores data tables, not models or agents; AI assets are catalogued in SageMaker (docs).

• AI Lineage: No lineage in the Glue Data Catalog (docs).

• AI Regulatory Compliance: No AI regulatory templates (docs).

• AI Model Health Score: No model scoring (docs).

AWS Glue Data Catalog workflows

• REST / GraphQL API: The AWS Glue API covers databases, tables, partitions, statistics and the new asset, glossary and search operations (docs).

• MCP Server: The AWS Labs Data Processing MCP server exposes Glue Data Catalog databases, tables and partitions to AI agents (docs).

• Python SDK: boto3 provides a pip-installable Glue client for all Data Catalog operations (docs).

• Webhooks: Data Catalog database, table and partition changes are emitted as Amazon EventBridge events that can route to HTTP targets (docs).

• dbt Integration: dbt metadata (manifest, tests) isn't ingested into the Glue Data Catalog; the dbt-glue adapter only runs dbt models on Glue (docs).

• Browser Extension: No browser extension (docs).