🤖 MCP Enterprise AIOps Agent & Gateway

A Production-Grade Model Context Protocol (MCP) implementation in Python powered by Groq Llama-3.3-70B and Streamlit Web UI.

This project functions as an Autonomous Self-Healing AIOps Engine that monitors system metrics, detects memory/CPU threshold breaches, and dynamically executes remediation tools to restore server health without human intervention.

🌟 Key Features

• Autonomous LLM Agent: Leverages Anthropic's Model Context Protocol (MCP) and Groq LLM Function Calling.

• Self-Healing Architecture: Automatically detects high CPU/RAM usage (e.g. >80%) and triggers remediation scripts down to safe levels (~35%).

• Dynamic Tool Router: Centralized registry mapping LLM tool requests to python functions.

• Zero-Trust Security: Token authentication middleware and Human-in-the-Loop authorization hooks for sensitive operations.

• Multi-Interface Support: Operates via Streamlit Web UI & CLI Terminal with Winston-style Python logging.

Related MCP server: AIOps MCP

🛠️ Project Architecture

mcp-enterprise-aiops-agent/ ├── .env ├── .gitignore ├── requirements.txt ├── README.md ├── app_gui.py └── src/ ├── init.py ├── security/ │ ├── init.py │ ├── auth.py # Zero-Trust JWT/Token Validation │ └── privacy_gateway.py # Presidio-style PII Redaction & Data Masking ├── tools/ │ ├── init.py │ ├── system_tool.py # System Health Monitoring & Self-Healing │ └── external_api_tool.py# Microservice Tool Integration ├── core/ │ ├── init.py │ ├── policy_engine.py # OPA-style Policy Enforcement & Human-in-the-Loop │ ├── tool_router.py # Dynamic MCP Tool Registration & Execution │ └── llm_agent.py # Multi-turn Autonomous Agent Loop └── utils/ ├── init.py └── logger.py # Structured Audit Logging (OpenTelemetry style)vice Data)