Python 開發筆記 - 使用 Nginx / WSGI / Gunicorn / Flask 進行 Python API 服務的上線整合 @ Ubuntu

近期工作上在 node.js 服務上，多開了一個 python api 服務做整合應用，目前先試著混合架構來擠擠機器資源而不是 micorservice 架構。

實作就是 Nginx 擋在前面，接著有些 requests 交給 node.js 運作，有些 requests 交給 python api 服務，整體上就是透過 Proxy Pass 架構：

$ cat /etc/nginx/conf.d/service.conf | grep location

    location / {

    location /node/ { rewrite ^/node/(.*)$ /$1 break; proxy_set_header X-Real-IP $remote_addr; proxy_set_header X-Forwarded-For $proxy_add_x_forwarded_for; proxy_redirect off; proxy_http_version 1.1; proxy_read_timeout 60; proxy_pass http://localhost:3000; }

    location /python/ { rewrite ^/python/(.*)$ /$1 break; proxy_set_header X-Real-IP $remote_addr; proxy_set_header X-Forwarded-For $proxy_add_x_forwarded_for; proxy_redirect off; proxy_http_version 1.1; proxy_read_timeout 60; proxy_pass http://unix:/var/run/service-py.sock; }

其中 node.js 是跑在 3000 port 服務，而 python api 跑在 unix:/var/run/service-py.sock ，上述 Nginx 設定檔剛好可以作為筆記，屬於兩種不同的設計方式，此外，這邊收到 requests 導向到 node.js 或 python api 時，都會刻意再去掉一些 prefix ，讓後面的服務開發比較多彈性。

回到 python api ，此次採用 Flask framework，他的運行很簡單：

```

% cat app.py

...

if __name__ == '__main__':

    app.run(host='localhost', port=3001)

```

直接執行法：

% python3 app.py 

 * Serving Flask app 'app' (lazy loading)

 * Environment: production

   WARNING: This is a development server. Do not use it in a production deployment.

   Use a production WSGI server instead.

 * Debug mode: off

WARNING: This is a development server. Do not use it in a production deployment. Use a production WSGI server instead.

 * Running on http://localhost:3001

Press CTRL+C to quit

使用 Flask 指令執行：

$ FLASK_APP=app.py flask run --host 0.0.0.0 --port 3001

 * Serving Flask app 'app.py' (lazy loading)

 * Environment: production

   WARNING: This is a development server. Do not use it in a production deployment.

   Use a production WSGI server instead.

 * Debug mode: off

WARNING: This is a development server. Do not use it in a production deployment. Use a production WSGI server instead.

 * Running on all addresses (0.0.0.0)

 * Running on http://127.0.0.1:3001

 * Running on http://x.x.x.x:3001

Press CTRL+C to quit

最後則是為了穩定性，採用 Nginx + WSGI 整合方式，透過 Gunicorn 來工作：

$ cat wsgi.py 

from app import app

if __name__ == "__main__":

    app.run()

$ sudo gunicorn --workers 4 --bind unix:/var/run/service-py.sock -m 777 wsgi:app

[2024-08-07] [1701223] [INFO] Starting gunicorn 20.0.4

[2024-08-07] [1701223] [INFO] Listening at: unix:/var/run/service-py.sock (1701223)

[2024-08-07] [1701223] [INFO] Using worker: sync

[2024-08-07] [1701225] [INFO] Booting worker with pid: 1701225

[2024-08-07] [1701226] [INFO] Booting worker with pid: 1701226

[2024-08-07] [1701227] [INFO] Booting worker with pid: 1701227

[2024-08-07] [1701228] [INFO] Booting worker with pid: 1701228

如果要把運行也包裝系統指令方便處理，就只需：

$ cat /etc/systemd/system/my-py.service 

[Unit]

Description=my-py-service

After=network.target

[Service]

Type=simple

User=root

WorkingDirectory=/path/project

#ExecStart=/usr/bin/python3 app.py

#Environment=FLASK_APP=app.py

#ExecStart=/usr/bin/python3 -m flask run --port 3001 --host 0.0.0.0

ExecStart=/usr/bin/gunicorn --workers 4 --bind unix:/var/run/service-py.sock -m 777 wsgi:app

Restart=on-failure

[Install]

WantedBy=multi-user.target

後續就可以透過以下方式管理：

$ sudo systemctl status my-py.service

$ sudo systemctl stop my-py.service

$ sudo systemctl start my-py.service

$ sudo systemctl restart my-py.service