End-to-End Tests¶
Validate complete workflows from start to finish, replicating real conditions.
Focus¶
- Critical user journeys
- Full pipeline runs
- Production-representative scenarios
Guidance¶
Use selectively. End-to-end tests are expensive to write, slow to run, and brittle to maintain. Reserve them for the workflows where failure would have the highest business impact.
- Run in a UAT or staging environment
- Use production-scale data volumes where possible
- Do not use as a substitute for unit or integration tests
Data-Specific Examples¶
- Full dbt model runs in UAT
- Complete data pipeline from ingestion to final table
- Data quality checks across the full dataset
- Environment-specific validation before promotion to production