title	API Reliability RL Environment
emoji	🚀
colorFrom	purple
colorTo	blue
sdk	docker
app_file	app.py
pinned	false

🚀 Cost-Aware API Reliability RL Environment

🧠 Overview

This project implements a real-world reinforcement learning (RL) environment that simulates API reliability challenges in backend systems.

Agents must make intelligent decisions under uncertainty to balance:

• ✅ Success rate
• ⏱️ Latency
• 💰 Cost

---

🎯 Objective

Enable agents to learn optimal strategies for handling unreliable APIs using the OpenEnv framework.

---

🧩 State Space (Observation)

Feature	Description
api_status	success / slow / failed
latency	Response time in ms
retry_count	Number of retries performed
api_cost	Cost of API usage
system_load	low / medium / high

---

⚡ Action Space

Action	Description
accept	Accept current API response
retry	Retry the same API
switch_api	Switch to backup API
use_cache	Use cached response (fast, cheap)
return_error	Stop and return failure

---

🏆 Reward Function

• +8 → successful response
• −0.02 × latency
• −5 × api_cost
• −2 × retry_count
• −8 → failure
• Bonus/penalty for decision quality

---

🧪 Tasks (Difficulty Levels)

Task	Description
Easy	Low failure probability (~30%)
Medium	Moderate failures (~50%)
Hard	High failure + cascading effects

---

📁 Project Structure

api-reliability-rl/
├── server/
│   ├── __init__.py        # Makes server a Python package
│   ├── app.py             # FastAPI server (OpenEnv)
│   └── environment.py     # RL Environment logic
├── app.py                 # Gradio UI
├── models.py              # Action, Observation, State models
├── inference.py           # Agent inference script
├── requirements.txt       # Python dependencies
├── openenv.yaml           # OpenEnv config
├── Dockerfile             # Docker deployment
└── README.md

---

📦 Local Setup

pip install -r requirements.txt

# Terminal 1 - Start FastAPI server
uvicorn server.app:app --host 0.0.0.0 --port 8000

# Terminal 2 - Start Gradio UI
python app.py

---

🧪 Inference Script

python inference.py

---

🐳 Docker Setup

docker build -t api-env .
docker run -p 8000:8000 -p 7860:7860 api-env

---

🌐 API Endpoints

Endpoint	Method	Description
/reset	POST	Reset environment
/step	POST	Take an action
/state	GET	Get current state

---

🛠️ Tech Stack

• OpenEnv
• FastAPI
• Gradio
• Docker
• Hugging Face Spaces
• OpenAI-compatible API (Qwen via HF Router)

---

👥 Team

• PALETI SAI TARUN — saitarunpaleti@gmail.com
• Rahil Ahmed — rahilahmed1305@gmail.com
• Ganesh Rayapati — nehapavanr@gmail.com