🧠 NeuroOps Backend

AI-Powered DevOps Intelligence Platform — Backend Microservices

---

📖 Overview

NeuroOps is an AI-powered DevOps intelligence platform that automatically analyzes GitHub Pull Requests and detects potential risks before code is merged.

The backend system processes GitHub webhook events, performs AI-based code analysis, stores insights, and exposes APIs used by the NeuroOps analytics dashboard. This repository contains the event-driven backend architecture responsible for:

• Pull request ingestion
• AI risk analysis
• GitHub comment automation
• Analytics aggregation
• Real-time frontend updates

---

🏗 Architecture

NeuroOps uses an event-driven microservice architecture to decouple webhook ingestion from AI processing.

graph TD
A[GitHub Pull Request] --> B[GitHub Webhook Service]
B --> C[Redis Queue - BullMQ]
C --> D[Worker Service]
D -->|AI Analysis| E[NVIDIA Kimi 2.5 API]
D --> F[(MongoDB Atlas)]
F --> G[Gateway API]
G --> H[React Dashboard]
G -->|SSE| H

Loading

---

🧩 Core Services

🔌 GitHub Service

Receives webhook events from GitHub when pull requests are opened or updated.

• Responsibilities: Verify webhook signatures & push PR jobs into a Redis queue.

⚙️ Worker Service

Processes queued pull request events.

• Responsibilities: Fetch PR diff from GitHub, perform AI analysis, extract risk scores, store results in MongoDB, post AI review comments, and notify the dashboard via SSE.

⛩️ Gateway Service

REST API layer used by the frontend dashboard.

• Responsibilities: Serve PR analytics, expose REST endpoints, and broadcast SSE updates to the dashboard.

---

✨ Key Features

• 🤖 AI Pull Request Risk Analysis: Each pull request diff is analyzed using the NVIDIA Kimi-2.5 AI model to detect potential bugs, security risks, performance issues, and risky patterns.

• 💬 Automated GitHub PR Comments: AI feedback is automatically posted directly to the pull request as a review comment, streamlining the peer review process.

• ⚡ Real-Time Dashboard Updates: The frontend receives updates instantly using Server-Sent Events (SSE) the moment an analysis is finalized.

• 📦 Queue-Based Processing: Powered by BullMQ + Redis to ensure reliable background job execution and seamless horizontal scalability.back is automatically posted directly to the pull request as a review comment.

---

🛠 Tech Stack

Component	Technology
Backend	Node.js, Express
Queue	BullMQ, Redis (Upstash)
Database	MongoDB Atlas
AI	NVIDIA Kimi-2.5 API
Infrastructure	Docker, Render

---

📂 Project Structure

.
├── services/
│   ├── github-service/    # Webhook handler & Job producer
│   ├── worker-service/    # AI Analysis & PR Reviewer
│   └── gateway-service/   # API & Dashboard SSE updates
├── docker-compose.yml     # Local orchestration
├── .env.example           # Template for environment variables
└── README.md              # Project documentation

---

⚙️ Environment Variables

Create a .env file in the root with the following keys:

# GitHub Configuration
GITHUB_TOKEN=your_github_token
WEBHOOK_SECRET=your_webhook_secret

# Database & Queue
MONGO_URI=mongodb+srv://user:password@cluster.mongodb.net/neuroops
REDIS_URL=rediss://default:password@upstash-url:6379

# AI Configuration (NVIDIA Kimi-2.5)
NVIDIA_API_KEY=your_kimi_api_key

# Service URLs
GATEWAY_URL=[https://neuroops-gateway.onrender.com/](https://neuroops-gateway.onrender.com/)

---

💻 Running Locally

1. Clone the repository

   git clone [https://github.com/dcpro8/neuroops-backend.git](https://github.com/dcpro8/neuroops-backend.git)
   cd neuroops-backend

2. Install dependencies

   npm install

3. Run services

   docker compose up --build

---

🚢 Deployment

The backend is deployed on Render using a microservices architecture.

• Services Deployed: github-service,worker-service,gateway-service.

• External Cloud Services:

  • Database: MongoDB Atlas

  • Queue: Upstash Redis

  • AI Engine: NVIDIA Kimi-2.5 API

---

🛣 API Endpoints

Method	Endpoint	Description
POST	/webhook	Receive GitHub webhook events
GET	/api/prs	Return analyzed pull requests
GET	/api/analytics/*	Analytics endpoints used by dashboard
GET	/api/events	SSE stream for real-time updates

---

👤 Author

Dhruv Chauhan

📄 License

This project is licensed under the MIT License.