Question
Design a data-quality gating system for a daily partitioned warehouse pipeline where bad upstream data (a source bug dumping nulls, a 10x row-count drop, a timezone shift, a duplicate load) currently silently corrupts a finance fact table that ~40 downstream dashboards and an ML feature pipeline depend on. The goal: a bad partition must be detected and BLOCKED from being published/consumed before it poisons downstream, while a healthy partition flows through automatically with no human in the loop. False positives that block good partitions are nearly as costly as missed bad ones.
Clarify scale and constraints first. Propose a clean component breakdown, then go deep on the hard parts — data model, bottlenecks, consistency, failure modes — and name the trade-offs you are making.