Governed Self-Service Data Quality Gates for Cloud Migration: Distribution- and Model-Aware Cutover Controls

Abstract

Self-service data quality tools allow the lay user to run their enterprise data quality checks as part of the cloud migration‚ but usually as just one fixed rule‚ run against a fixed baseline․ I argue that the self-service gate should only be trusted for a cutover decision if it detects data distribution shift‚ and also takes the importance of the changed feature to the model downstream into account․ Supported by practitioner experience with migration validation tooling and the drift detection‚ feature attribution‚ and self-service governance literatures‚ I describe a model of what the gate needs beyond accessibility (to be distribution-aware and model-aware) and who is responsible for a cutover decision that the non-specialist-run gate has supported (the team‚ or the organization)․