Running an ML Visualization Dashboard from a Docker in Azure Web App

Question

HI -- I have an app.py, requirements.txt, and dockerfile that are all loaded within Azure ML and aligned with an ACR registry. I've also created a web app and linked the docker to the web app. The problem is that when I try to access the published web app site, it does not show the interactive dashboard as specified in my app.py code. It instead shows a blank webpage that has an "ML Model Prediction" and a blank box to enter a value. How do I get the web app to show the actual code from the flask app in my app.py. The app works fine in test environment.

Answer 1

Hello @Ronald Iammartino !

Welcome to Microft QnA!

The problem seems to be in the Web App deployment

If you have aby .env values , you have to add them to APP Configuration Settings

Enable Logging and provide the errors

Sometimes Packages are not compatible and you have to find the correct versions

How is your Dockerfile ? Can you post it please ? Same for app.py if possible !

---

I hope this helps!

Kindly mark the answer as Accepted and Upvote in case it helped!

Regards