Observability AIOps using Isolation Forest and LLaMA with FAISS for RAG

Overview

This project implements anomaly detection using the Isolation Forest model and integrates LLaMA with FAISS for Retrieval-Augmented Generation (RAG) to create cases based on detected anomalies. The solution enables automated anomaly detection and root cause analysis using AI-driven observability.

Features

Anomaly Detection: Uses Isolation Forest for unsupervised anomaly detection.

Case Creation: LLaMA generates cases based on detected anomalies.

Efficient Retrieval: FAISS provides fast similarity search for historical anomaly cases.

Scalable: Supports large datasets and real-time monitoring.

Installation

Prerequisites

Python 3.8+

pip

PyTorch (for LLaMA)

FAISS