Fix the problem of no listening logs when deploying a web UI service created using gradio in a Kubernetes environment

Problem Description

When deploying a web UI service created using gradio in a Kubernetes environment, there may be no listening logs in the Pod container, and sometimes there may be a blockage from the beginning, resulting in the inability to provide HTTP services externally.

Sample Code:

import gradio as gr

def greet(name):
    return "Hello " + name + "!"

demo = gr.Interface(fn=greet, inputs="textbox", outputs="textbox")

if __name__ == "__main__":


The reason for the lack of listening logs is that the gradio code uses print() to output the listening logs, and the default output of print() uses a buffer. In container environments like Kubernetes, there may be a delay in displaying the output of print().


This problem can be solved using any of the following methods.

Set Environment Variables

You can solve this problem by setting the value of the environment variable PYTHONUNBUFFERED to a non-empty string:

  • You can set this environment variable in the Dockerfile:

  • You can also set this environment variable in the definition of the Pod:

       value: 'true'

Set Python Execution Parameters

You can achieve the same effect by adding the -u parameter when running the service code with python.

python -u

Configure Pod Definition to Use Standard Input

You can also solve this problem by configuring the container definition in the Pod to use standard input.

Sample Pod YAML:

 - image: xxx
   tty: true
   stdin: true