Amazon SageMaker Unified Studio now supports data quality rule authoring and evaluation
Define and enforce data quality rules on catalog tables and ETL pipelines directly in SageMaker Unified Studio — no separate Glue console needed.
View original announcement →Visual Summary
graph TD
A{{SageMaker Unified Studio Data Quality}}:::announced
B((Data Practitioners)):::external
C(AWS Glue Data Quality):::compute
D([DQDL Rule Authoring]):::feature
E([Data at Rest Evaluation]):::feature
F([Data in Transit Evaluation]):::feature
G(Data Catalog Tables):::storage
H(Visual ETL Jobs):::compute
I(Data Lakes):::storage
B ==>|"authors rules"| A
A ==>|"powered by"| C
A -->|"defines"| D
D -->|"applies to"| E
D -->|"applies to"| F
E -->|"evaluates"| G
F -->|"validates in"| H
H -->|"writes to"| I
G -.->|"scheduled checks"| C
C -->|"returns results"| A
classDef announced fill:#ff9900,stroke:#ec7211,color:#fff,font-weight:bold
classDef compute fill:#e3f2fd,stroke:#1565c0,color:#1565c0
classDef storage fill:#e8f5e9,stroke:#2e7d32,color:#2e7d32
classDef feature fill:#fff3e0,stroke:#e65100,color:#e65100
classDef external fill:#f5f5f5,stroke:#616161,color:#616161
What's New
Amazon SageMaker Unified Studio now integrates AWS Glue Data Quality capabilities, enabling data engineers, analysts, and data scientists to author, evaluate, and review data quality rules without leaving the unified workspace. The feature supports two distinct workflows: quality checks on data at rest via catalog table assets, and inline quality evaluation within Visual ETL jobs for data in transit. Rules are authored using the Data Quality Definition Language (DQDL), the same DSL used natively in AWS Glue Data Quality, covering dimensions such as completeness, uniqueness, freshness, and accuracy.
How It Works
- DQDL Rule Authoring: Users write rulesets using DQDL, which provides 31 built-in rule types covering data quality dimensions like completeness, uniqueness, freshness, and accuracy — the same language and syntax used in standalone AWS Glue Data Quality.
- Data at Rest (Catalog Assets): A dedicated "Data Quality" tab is surfaced on any catalog table asset within SageMaker Unified Studio, where users can define rulesets, trigger on-demand evaluations, schedule recurring evaluations, and inspect per-rule pass/fail results and quality scores over time.
- Data in Transit (Visual ETL Jobs): Users can insert an "Evaluate Data Quality" transform node into any Visual ETL pipeline; quality results are captured and surfaced as part of the job run details, allowing inline validation before data reaches downstream targets.
- Powered by AWS Glue Data Quality: The underlying evaluation engine is AWS Glue Data Quality, meaning compute, rule execution, and result storage leverage the same managed infrastructure — no separate pipeline or service configuration is required.
- Identity and Access: The feature works in both AWS IAM Identity Center-based and IAM-based SageMaker Unified Studio domains, inheriting existing access control patterns.
Why It's Important
- Shift-Left Data Quality: By embedding quality checks directly into the data catalog and ETL authoring experience, teams can catch and remediate data issues before bad data propagates into data lakes, BI dashboards, or ML training pipelines — reducing costly downstream failures.
- Reduced Context Switching: Previously, data practitioners had to navigate to the AWS Glue console or build separate pipelines to run quality evaluations; this integration eliminates that friction by centralizing the workflow inside SageMaker Unified Studio.
- Proactive ML Data Governance: Data scientists can validate training and feature datasets against defined quality thresholds directly from the catalog, improving model reliability and reducing silent data drift issues.
- Operational Monitoring Without Pipelines: The scheduled evaluation capability on catalog assets means teams can continuously monitor data quality over time without engineering a dedicated orchestration pipeline.
- Consistent Quality Standards: Using the same DQDL language across Glue and SageMaker Unified Studio ensures that quality rules are portable and consistent across teams and tooling.
How It's Different
- Unified Surface vs. Separate Console: Unlike using AWS Glue Data Quality directly in the Glue console, this integration brings rule authoring and results into SageMaker Unified Studio alongside data discovery, ETL, and ML workflows — no context switching required.
- Catalog-Native Quality Tab: The dedicated Data Quality tab on catalog assets is a new UX pattern that ties quality scores directly to data assets, making quality status visible at the point of data discovery rather than buried in a separate monitoring tool.
- ETL-Inline Evaluation: The "Evaluate Data Quality" transform in Visual ETL allows quality gates to be embedded as first-class nodes in pipeline logic, rather than as post-hoc checks or external validation scripts.
- No Additional Rule Language to Learn: Teams already using AWS Glue Data Quality can reuse existing DQDL rulesets and knowledge directly, lowering the adoption barrier compared to learning a proprietary quality framework.
- Dual Workflow Coverage: Most data quality tools focus on either static dataset profiling or pipeline-level checks; this feature explicitly addresses both data at rest and data in transit within a single integrated experience.
When to Prefer It
- When your team already uses SageMaker Unified Studio as the primary workspace for data engineering and ML, and you want to avoid managing quality checks in a separate AWS console.
- When you need to validate catalog datasets before ML training, ensuring that feature tables or raw datasets meet defined quality thresholds (e.g., completeness > 95%) prior to model development.
- When building Visual ETL pipelines that feed data lakes or warehouses, and you want to enforce quality gates inline to prevent bad records from landing in downstream storage.
- When you need scheduled, ongoing data quality monitoring on catalog tables without building and maintaining a separate orchestration workflow (e.g., Step Functions or Glue Workflows).
- When your organization is standardizing on DQDL for data quality rules and wants consistent rule definitions shared across Glue jobs and SageMaker Unified Studio workflows.
- When data governance requirements mandate documented quality scores on data assets, as the catalog-level quality tab provides persistent, auditable pass/fail history tied to specific tables.
Availability
- GA Status: Generally available as of May 20, 2026.
- Regional Availability: Available in all AWS Regions where Amazon SageMaker Unified Studio is currently supported.
- Domain Compatibility: Supported in both AWS IAM Identity Center-based and IAM-based SageMaker Unified Studio domains.
- Underlying Engine: Powered by AWS Glue Data Quality; pricing is expected to follow AWS Glue Data Quality consumption-based pricing (Data Quality Units / DQUs), though users should consult the AWS pricing page for exact details.
- Rule Types: 31 built-in DQDL rule types are available, covering completeness, uniqueness, freshness, accuracy, and other standard data quality dimensions.
- Documentation: Full details available at the Amazon SageMaker Unified Studio User Guide (data-quality-catalog page) and the AWS Glue DQDL reference documentation.
