Awesome AI for Infra › Root Cause Analysis

cuebook/CueObserve

⭐ 234 Python added to this list on 2026-06-14 repository created 2021-06-22

CueObserve provides a framework for monitoring business and operational metrics by detecting anomalies in time-series data sourced from SQL data warehouses and databases. The platform integrates directly with various data sources like Snowflake, BigQuery, Redshift, and Postgres, allowing users to define datasets using SQL `GROUP BY` queries. It then automatically transforms the SQL query results into time-series data. CueObserve employs forecasting models, primarily Facebook's Prophet, to predict expected metric behavior and identify significant deviations as anomalies. Upon detecting an anomaly, it offers a one-click root cause analysis feature, helping users understand why a metric is not performing as expected by splitting the anomalous metric by different dimensions. The system is designed to be configurable, enabling users to define multiple anomaly detection jobs on a single dataset, specify aggregation levels, and configure alerts. It includes an in-built scheduler for recurring analysis and integrates with communication tools like Slack for anomaly notifications. While powerful for historical and near real-time analysis, it is explicitly not designed for streaming data or real-time anomaly detection. CueObserve aims to provide clear insights into metric performance and underlying issues, making it a valuable tool for data-driven operations and business intelligence.

https://github.com/cuebook/CueObserve

anomaly detectionroot cause analysistime-seriesdata warehouseSQLAIOpsmetrics monitoringpredictive analytics