# Not every quality check should stop the pipeline Canonical URL: Domain: Pipeline Design ยท Difficulty: medium ## Problem Not every quality check should stop the pipeline. The section's rule is per-check: block when running is worse than not running, warn when running is better than not running but still imperfect. Five candidate checks sit on the canvas. A primary-key uniqueness violation, a required-column null spike above 5 percent, and a row count below 50 percent of baseline are blockers (downstream joins are now wrong; the consumer would be misled). A row count at 80 to 90 percent of average and a slight freshness drift are warnings. Pick warn or block by adding an authority annotation to each check's name and routing block-class checks to a paging destination. ## Related - [All practice problems](https://datadriven.io/problems) - [Mock interview mode](https://datadriven.io/interview/not-every-quality-check-should-stop-the-pipeline-the-sectio-c6e4924e) - [System Design Interview Questions](https://datadriven.io/data-engineering-system-design) - [Data Engineering Interview Prep Guide](https://datadriven.io/data-engineer-interview-prep) - [Daily Challenge](https://datadriven.io/daily) --- Source: DataDriven (https://datadriven.io). 100% free data engineering interview prep. Live code execution against Postgres 16, Python 3.11, and Spark sandboxes. No paywall, no premium tier, no signup gate.