🚀 Enterprise RAG Agent – LangGraph + FastAPI + Redis + Kubernetes

🧠 A production-grade, agentic AI backend that combines LangGraph, Retrieval-Augmented Generation (RAG), persistent memory via Redis, and secure FastAPI endpoints. Designed for full-scale deployment on Docker and Kubernetes with monitoring and API key security.

---

🎯 Objective

This project provides a modular boilerplate to build and deploy end-to-end agentic AI applications using LangGraph and RAG. It is production-ready, scalable, and extensible, enabling multi-turn conversations, external document integration, persistent memory, and comprehensive monitoring.

---

🧱 Architecture Overview

+------------------+       +-------------------+       +------------------+
|   Streamlit UI   | <---> | FastAPI Backend    | <---> |  LangGraph Agent |
+------------------+       +-------------------+       +------------------+
                                      |
                             +--------+--------+
                             |   Redis Memory   |
                             +--------+--------+
                                      |
                              +-------+--------+
                              | Vector Store    |
                              | (Chroma)  |
                              +----------------+

---

📁 Codebase Structure

/enterprise-rag-agent
│
├── app/
│   ├── main.py                 # FastAPI entrypoint
│   ├── config.py               # App-wide config (env vars, secrets)
│   ├── tools.py                # Tool functions used in LangGraph
│   ├── graph_manager.py        # LangGraph graph and flow logic
│   ├── vector_store/
│   │   ├── loader.py           # Load PDFs, markdown, etc.
│   │   ├── embedder.py         # Embedding logic using OpenAI or others
│   │   └── vectordb.py         # Vector DB logic (FAISS/Chroma)
│   ├── node_pool/
│   │   ├── retrieve_node.py
│   │   ├── generate_node.py
│   │   └── fallback_node.py
│   ├── state_manager/
│   │   └── memory_state.py     # Redis chat history threading
│   ├── integration/
│   │   └── confluence_api.py   # Placeholder for API integrations
│   ├── security/
│   │   └── auth.py             # API key validation
│   ├── monitoring/
│   │   └── metrics.py          # Prometheus-ready metrics
│   ├── error_handler.py        # Global exception handling
│
├── streamlit_ui/
│   └── ui.py                   # Multi-tab UI for user interaction
│
├── data/
│   └── sample_docs/            # PDFs, Markdown, CSVs for ingestion
│
├── docker/
│   └── Dockerfile              # Backend Docker setup
│
├── k8/
│   ├── deployment.yaml         # Kubernetes deployment
│   └── service.yaml            # Kubernetes service
│
├── .env                        # Secrets like API keys
├── .gitignore                 # Ignored files
├── requirements.txt           # Python dependencies
└── README.md                  # You’re reading it

---

⚙️ Setup Instructions

1. 🧬 Clone Repo

git clone https://github.com/your-org/enterprise-rag-agent.git
cd enterprise-rag-agent

2. 🔐 Create .env

AZURE_ENDPOINT = 
AZURE_API_KEY = 
OPENAI_API_TYPE = 
AZURE_DEPLOYMENT = 
AZURE_API_VERSION = 
AZURE_MODEL=
GEMINI_API_KEY=
LANGSMITH_TRACING=true
LANGSMITH_ENDPOINT=
LANGSMITH_API_KEY=
LANGSMITH_PROJECT=
SERPER_API_KEY=
RAG_API_KEY=

3. 📦 Install Dependencies

pip install -r requirements.txt

4. 🚀 Start Redis

docker run -d -p 6379:6379 redis

5. 🧠 Run FastAPI Backend

uvicorn app.main:app --reload

6. 🖥️ Run Streamlit UI

streamlit run streamlit_ui/ui.py

---

🔐 Security

• ✅ API key validation via X-API-Key
• ✅ Secure loading from .env using python-dotenv
• ✅ Centralized auth.py logic

---

🧠 Memory & Chat Threading

• 🔁 Redis-based persistent state across sessions
• 🧵 Threaded conversations
• 💾 State autosaved before/after LangGraph execution

---

🤖 LangGraph Agent

• Nodes:
  • 📥 Retrieve Node – RAG doc search
  • ✍️ Generate Node – LLM reasoning
  • 🧯 Fallback Node – edge case handler
• Built using LangGraph’s composable workflows

---

📡 API Reference

POST /ask

Ask a question to the agent.

POST /ask
Headers:
  X-API-Key: your_api_key

Body:
{
  "question": "What is the sales trend for Q1?"
}

Response

{
  "final_answer": "The sales trend for Q1 was an upward growth of 18%."
}

---

📊 Monitoring

• 📈 /metrics endpoint exposes Prometheus counters
• Tracks:
  • ⚡ Request duration
  • 🔢 Total requests
  • ✅ Response status codes
  • 🔍 RAG operation metrics

---

🐳 Docker Support

docker network create rag-net
 docker run -d   --name redis   --network rag-network   redis:latest
docker network connect rag-net redis

 docker run -d   --name rag-backend   --network rag-net   -p 8000:8000   rag-backend:latest

docker run -d   --name rag-frontend   --network rag-net   -p 8501:8501   rag-frontend:latest<img width="902" height="189" alt="image" src="https://github.com/user-attachments/assets/66ef1b9e-081d-4c5c-9b61-93f61d7576c5" />

---

☸️ Kubernetes Deployment

kubectl apply -f deployment/

---

✅ Feature Highlights

• 🔐 Secure API layer
• 📚 Vector Search with FAISS/Chroma
• 🔄 Persistent Redis memory
• 🧠 LangGraph agent orchestration
• 🧪 Monitoring with Prometheus
• 🖥️ Multi-tab Streamlit UI
• 🐳 Docker-ready
• ☸️ Kubernetes-ready

---

🚧 Coming Soon

• 🚀 CI/CD via GitHub Actions
• ☁️ Azure Blob, SharePoint, Confluence integrations
• 💬 Slack & Teams notifications
• 🧾 Session transcripts and summaries

---

👨‍💻 Maintainer

Pradeep P. – Software Professional | AI & Cloud Engineer

Feel free to fork, contribute, or raise issues!

---