DEMO

PyCon TW 2024 talk Using Load Testing to Unveil Performance Differences Between Sync and Async: Web Server as a Case Study

---

Projects

account_password_management

• Load Testing Tool : Locust
• Backend Framework : FastAPI (async)
• ASGI Web Server : Uvicorn
• Database : PostgreSQL
• ORM : SQLAlchemy
• Load Balance : Nginx
• lint : ruff、isort

Requirements

• python 3.11+
• poetry

Installation

1. Clone the repo

   git clone

2. Install Python packages

   poetry install

3. Install PostgreSQL

4. Set PostgreSQL account

5. Create DB and Schema

6. Create .env

   .env example

   STAGE="prod"
   SERVICE_PORT=3000
   WORKERS=1
   HOST="0.0.0.0"
   
   DB_HOST="localhost"
   DB_PORT="5432"
   DB_USER="postgres"
   DB_PASSWORD="password"
   DB_DATABASE="load_testing"
   DB_POOL_SIZE=10
   DB_MAX_OVERFLOW=5
   DB_POOL_RECYCLE=300

Usage

• Startup the services

  python app.py

• Execute load testing

  locust --config .\locust_files\demo_config.conf {User Class}

API Document

1. Startup the services
2. Open a browser to http://localhost:3000/docs#

---

Contact

TingYi Xiao - [email protected]

---

Acknowledgments

• Locust
• FastAPI
• Uvicorn
• Poetry
• SQLAlchemy
• PostgreSQL
• Nginx
• README