DriftLedger

DriftLedger is a web app (with optional CLI agent) that monitors “silent” changes in big-data pipelines: schema shifts, distribution drift, null spikes, late-arriving data, and metric breakages. Instead of trying to replace your warehouse/lakehouse, it plugs into what you already run (Snowflake/BigQuery/Databricks) and continuously profiles key tables and critical metrics. It creates baselines, detects anomalies, and routes actionable alerts to Slack/Jira with a clear root-cause trail: which upstream table changed, when it started, and which downstream models/dashboards are impacted. The MVP focuses on a small set of high-value tables and metrics, because broad coverage is expensive and noisy. This is a combination traditional + AI app: traditional rules and statistical tests for detection, plus AI to summarize incidents, suggest likely causes, and draft remediation steps and stakeholder updates.