NovaOps AI

AI-powered incident analysis and remediation system for operational logs

NovaOps is an end-to-end system that ingests raw system logs and produces structured incident analysis, including severity classification, root-cause hypotheses, remediation plans, and validation checks.

It combines LLM reasoning + vector retrieval + structured outputs to simulate real-world incident debugging workflows used in DevOps and SRE environments.

---

✨ Features

• 🔍 Log Analysis — Paste raw logs and extract meaningful signals
• 🧠 AI Reasoning Engine — Structured triage and diagnosis using LLMs
• 📊 Incident Console — Predefined incident simulations (autoscaling, rollout, CPU spike)
• 🧩 Evidence Retrieval — FAISS-based vector search for contextual grounding
• 🛠️ Remediation Planning — Step-by-step actionable recovery plan
• ✅ Validation Checks — Ensures correctness of applied fixes
• ⚠️ Risk Awareness — Safety considerations before applying changes
• 📁 History Tracking — Persisted analysis results

---

🏗️ Architecture

User Input (Logs / Incident)
        ↓
Embedding + Retrieval (FAISS)
        ↓
LLM Reasoning (Triage)
        ↓
Hypothesis Generation
        ↓
Remediation Planning
        ↓
Validation + Reporting

---

🛠️ Tech Stack

Frontend

• Next.js (App Router)
• TypeScript
• Tailwind CSS

Backend

• FastAPI
• FAISS (vector search)
• Pydantic
• AWS Bedrock (Amazon Nova)

---

📸 Screens

Analyze Page

Paste logs → get structured output:

• Summary
• Severity
• Root cause
• Plan
• Validation

Incident Console

Predefined real-world scenarios:

• Autoscaling misconfiguration
• Bad deployment rollout
• CPU spike

---

🚀 Getting Started

1. Clone repository

git clone https://github.com/YOUR_USERNAME/novaops-ai.git
cd novaops-ai

---

2. Backend setup

cd backend
python -m venv .venv
.venv\Scripts\activate   # Windows
pip install -r requirements.txt
uvicorn app.main:app --reload

---

3. Frontend setup

cd frontend
npm install
npm run dev

---

4. Run the app:

👉 http://localhost:3000

---

🧪 Sample Logs (for testing)

Paste this into Analyze page:

2026-04-11T09:10:02Z INFO  auth-service  Login attempt user=admin ip=192.168.1.45
2026-04-11T09:10:05Z WARN  auth-service  Failed login attempt user=admin ip=192.168.1.45
2026-04-11T09:10:07Z WARN  auth-service  Failed login attempt user=admin ip=192.168.1.45
2026-04-11T09:10:10Z ERROR auth-service  Account locked due to repeated failures user=admin

2026-04-11T09:11:30Z ERROR db-service    Query timeout query_id=q-9123 duration=5000ms
2026-04-11T09:11:30Z ERROR api-gateway   Request failed route=/api/v1/orders status=504

2026-04-11T09:12:20Z ERROR order-service Exception in OrderProcessor
java.lang.NullPointerException: Cannot read property 'price' of undefined

---

🎯 Use Cases

• DevOps incident debugging
• SRE workflows
• Log anomaly detection
• Root cause analysis
• AI-assisted operations

---

🧠 Key Highlights

• Combines retrieval + reasoning + planning
• Produces structured outputs (not raw LLM text)
• Designed as a multi-stage AI system
• Simulates real-world production incident workflows

---

📌 Future Improvements

• Multi-agent orchestration (planner, validator agents)
• Real-time log streaming
• Incident clustering
• Slack / PagerDuty integration
• Cloud deployment (Render + Vercel)

---

👩‍💻 Author

Asmita Sonavane

---