TestGen connects to Redshift directly. This page picks a tool; the TestGen walkthrough for Redshift covers grants, profiling, and monitors. Running Redshift beside Aurora or S3 files? See the AWS answer.
What Redshift gives you natively
Constraints, and only one of them bites. Per AWS’s documentation, uniqueness, primary key, and foreign key constraints are informational. A load that violates them succeeds. The planner still trusts them, so a duplicate key can make SELECT DISTINCT return duplicates. Redshift enforces NOT NULL. Nothing in Redshift profiles a column, generates a test, or tells you a table stopped loading.
The AWS answer to that gap is AWS Glue Data Quality. It evaluates Redshift tables cataloged in Glue, recommends rules from table statistics, and runs ML anomaly detection. It runs on Glue workers, not as SQL in your warehouse, and you write the rules in DQDL.
Tools worth considering on Redshift
| DataOps TestGen | AWS Glue Data Quality | Monte Carlo | Soda | |
|---|---|---|---|---|
| Profiling | 55 characteristics per column, 32 hygiene detectors | Statistics through analyzers | Metadata and query logs | Dataset profiling in Soda Cloud |
| Generated data tests | 34 generated test types | Rule recommendations from statistics | Freshness and volume on by default | Contracts generated from schema or by AI |
| Observability monitors | Freshness, volume, schema, custom metrics | ML anomaly detection, dynamic rules | Freshness, volume, schema, metric, custom SQL | Metric monitors with anomaly detection |
| Data quality dashboards | Scorecards, CDE score, issue report | Data quality score per ruleset | Hosted UI | Hosted UI |
| Runs in Redshift | Yes, read-only SQL | No, on Glue workers | Monitor queries from Monte Carlo’s cloud or an agent | Scans run by a Soda-hosted or self-hosted runner |
| Open source | Yes, Apache 2.0 | Built on open-source Deequ | No | Soda Core, Elastic License 2.0 |
| Price basis | Free; Enterprise $100 per user plus $100 per connection per month | $0.44 per DPU-hour in AWS examples | Credits, paid per monitor | Free tier; Team $750 per month |
Sources, checked October 21, 2026: Glue Data Quality and Glue pricing; Monte Carlo Redshift and external tables; Soda’s Redshift connection and MCP tools list. Prices: Monte Carlo pricing and Soda pricing. DataKitchen’s Monte Carlo and Soda comparison pages carry annual cost estimates.
Who should pick which
Pick Glue Data Quality if your team lives in the AWS console, catalogs everything in Glue, and wants pay-per-run billing with ML anomaly detection built in. Pick Monte Carlo if you want a managed service that learns thresholds on its own and draws lineage from Redshift’s query logs, and the budget is there. Pick Soda if you run on data contracts written in YAML.
Pick TestGen if you want coverage without writing rules first. It profiles every column, generates tests from the profile, and runs them as SQL inside Redshift with only USAGE and SELECT grants. It also reads Spectrum external tables, and Redshift has been on its list since the open-source launch. The trade: you host and upgrade it yourself, and it draws no lineage from query logs. Compare it with Glue on the comparison page.
The core profiling, testing, and monitoring engine is free and open source. Enterprise adds team access, security, Commercial Pharma tests, source-data controls for AI, and support.
Documentation: AWS Redshift quickstart, Connect your Data, Access Requirements