Customer 360 & Real-Time Risk Intelligence Platform · View 09 of 20

Data Quality Framework

Rule definition, structural and semantic checks, disposition of failures, and the escalation and replay path back into the pipeline.

Editable source SVG draw.io All views

One rule set, two engines

The same declarative rule definitions run in-stream and post-load
Rules are versioned in Git and released through the same CI/CD gates as code
Stewards own thresholds; engineers own execution

Disposition policy

Row-level failure quarantines the row with a reason code; the batch continues
Dataset-level failure trips a circuit breaker and blocks downstream publication
Nothing is silently corrected — every fix is an auditable reprocessing event

Measurement

Daily quality scorecard per domain, trended against a 30-day baseline
Freshness SLA measured per dataset and reported alongside correctness
Quality metrics are published as events, so they are observable like any other signal