incident-root-cause-analyzer

Root-cause analysis of cascading failures uses logs, metrics, and traces to map temporal causality across microservices. This Skill automates anomaly detection, hypothesis testing, and evidence correlation to identify resource contention, backpressure propagation, and failure sequences, then generates structured reports with actionable remediation guidance.

Yes. The Skill accepts metrics data as CSV files in a directory alongside an incident time window. It processes the data using polars and numpy to score anomalies, correlate timeline events, and produce evidence artifacts including visualizations and a formal Root Cause Analysis report.

Reports include a structured Root Cause Analysis document, evidence charts generated with matplotlib, a Mermaid fault evolution diagram showing failure sequences, and an evidence index mapping findings to remediation actions.

Yes. The Skill accepts optional logs and traces directories alongside metrics data. It correlates all three data sources to verify temporal causality, rank anomalies, and test hypotheses across distributed system boundaries.

By automating data collection, anomaly scoring, and hypothesis evaluation across microservice boundaries, the Skill accelerates incident investigation. It enables faster identification of resource contention and backpressure-driven root causes compared to manual log analysis.

You need metrics data as CSV files, an incident time window (e.g., 07:00:00 ± 30s), and optionally logs or traces directories. The Skill handles data ingestion and processing automatically using polars, numpy, and matplotlib.