Skip to content

RajatRawal-06/Root-Cause-Agent

Folders and files

NameName
Last commit message
Last commit date

Latest commit

Β 

History

8 Commits
Β 
Β 
Β 
Β 

Repository files navigation

AI Incident Root Cause Analyzer

Welcome to the Root-Cause-Agent repository!

This repository hosts a professional, competition-ready incident investigation console engineered for modern SRE and DevOps teams. Designed specifically for the Agentic Hackathon, this project demonstrates a highly capable AI-driven workflow that safely evaluates active incidents, determines root causes through evidence chains, and suggests idempotent remediation actions.


🌟 Why This Project?

Modern infrastructure is complex. When a SEV-1 incident strikes, SREs spend valuable minutes digging through logs, correlating metrics across disparate services, and manually checking deployment histories.

The Root Cause Analyzer changes this paradigm.

Instead of manual hunting, our system deploys a swarm of specialized AI agents:

  1. Alert Intake Agent: Triage and assess incoming alerts.
  2. Logs Intelligence Agent: Parse and extract anomalies from distributed log streams.
  3. Metrics Analysis Agent: Identify latency spikes, error rate climbs, and resource exhaustion.
  4. Distributed Trace Agent: Follow requests across microservices.
  5. Deployment Analysis Agent: Correlate failures with recent code rollouts.
  6. Correlation Engine & Root Cause Analysis Agents: Synthesize findings into actionable evidence chains and assign confidence scores to hypotheses.

🎯 Key Features & Highlights

  • Explainable AI (XAI): We don't just output a guess. Every AI conclusion comes with a mapped evidence chain and a calculated confidence score, ensuring transparency for critical operational decisions.
  • Idempotent Remediation Engine: AI shouldn't break production. Our platform uses an audit-driven approach to simulate recovery commands (like rolling back deployments) safely, recording every action in an immutable audit log.
  • Real-time WebSocket Updates: A lively, responsive UI that pushes state changes immediately. No polling required.
  • Production-Oriented Architecture: Built with lightweight boundaries (SQLite, native Python API) but designed for a direct upgrade path to PostgreSQL, Kafka, and OpenTelemetry for enterprise scale.
  • WCAG-Compliant Professional Interface: A beautiful, minimalist "light mode" interface modeled after industry-standard tools like Grafana and Datadog, ensuring it feels instantly familiar to operations personnel.

πŸ“‚ Project Structure

All the core project files, documentation, and source code are located in the agenti-hackathon directory.

Inside the agenti-hackathon folder, you will find:

  • backend/: The core API server, Multi-Agent workflow logic, and data store models.
  • frontend/: The lightweight, accessible HTML/CSS/JS user interface.
  • tests/: Edge-case and workflow validation suites.
  • scripts/: Helpful deployment utilities.
  • README.md: Complete setup instructions, screenshots, and detailed API documentation.

πŸ‘‰ Navigate to the agenti-hackathon README for full details on how to run the project locally or via Docker.


Built with passion for the Agentic Hackathon. Empowering SREs with Agentic Intelligence.

About

AI-driven workflow that safely evaluates active incidents, determines root causes through evidence chains, and suggests idempotent remediation actions.

Resources

Stars

Watchers

Forks

Releases

Packages

Contributors

Languages