WebApp deployment keeps failing. What is the cause?

Anonymous

Greetings,
I'm trying to deploy my first WebApp from a GitHub repo of python code.
Very briefly, the repo has this structure:
root/
---- downloads <-- will contain the image uploaded by the user
---- templates/main.html <-- the html page that should be loaded
---- static/Images/ <-- now empy, will eventually contain the images for the page
---- lib/.... <-- stuff that the app will do once an image is loaded
---- application.py <-- Very simple python script to call the html page and route some functions to it
---- requirements.txt <-- the python requirements that will be installed during the build step
Of course, I've tried it multiple times locally, and it works perfectly fine.
Once I ask GitHub to deploy, however, I get this:

When I check the specific step of the deployment which went wrong, I get this:

The Log Stream of the application is also attached.209113-defier-webapp-log.txt

I tried to ask support, however, in the support page for this particular problem, I got told that a health check triggered a response from the customer support, and that they solved the problem 3 hours prior the time I'm writing this post.

I'd be very thankful for anyone kind enough to help me understanding what's going on here.

Anonymous

2022-06-09T08:29:22.87+00:00

I've tried to deploy it using Azure Visual Studio tools, and the deployment was successful. However, the application is not accessible from the web.
I've attached the deployment log, as you can see there are no errors.
However in the web app page I get the following:

209719-deployment-log.txt

Via ssh I inspected the log streams in the webapp machine, and found some informative infos which I attach as well.
209766-2022-06-09-pl1sdlwk000fb3-default-docker.log
209710-docker-log.txt

In particular in the file 209766-2022-06-09-pl1sdlwk000fb3-default-docker.log contains the lines:
2022-06-09T08:15:10.822381261Z Failed to find attribute 'app' in 'app'.
2022-06-09T08:15:10.851232680Z [2022-06-09 08:15:10 +0000] [78] [INFO] Worker exiting (pid: 78)
2022-06-09T08:15:10.937709036Z [2022-06-09 08:15:10 +0000] [75] [INFO] Shutting down: Master
2022-06-09T08:15:10.937765036Z [2022-06-09 08:15:10 +0000] [75] [INFO] Reason: App failed to load.

This seems to be the only error in the startup.
Ryan Hill 26,146 Reputation points Microsoft Employee

2022-06-09T13:44:24.027+00:00

What's your project structure?

2 answers

Manu Philip 16,986 Reputation points MVP

2022-06-07T10:48:11.287+00:00
It sounds like the port configured is not able to reach by the deployment as indicated in the below log entry

Container defier_0_8d248356 didn't respond to HTTP pings on port: 8000, failing site start

Check, if you can change the port to Port 80 to resolve the situation

--please don't forget to upvote and Accept as answer if the reply is helpful--
Please sign in to rate this answer.

1 person found this answer helpful.
Anonymous

2022-06-07T16:00:21.433+00:00

Thanks for the reply @Manu Philip !
If I understood correctly, you suggested to change the port of the application to Port 80. What I'm wondering is:

I specify this in application.py (which means changing the line app.run() to app.run(port=80)

Do I have to do it somewhere else in Azure?

What I've tried after your answer:

Specify the port in application.py to 80

Specify the port in application.py to 8000

Both led to the same timeout error you highlighted in your answer.

Manu Philip 16,986 Reputation points MVP

2022-06-07T16:17:36.423+00:00

Usually, we expose the port needed. For example, if we go with Dockerfile, we should expose the port needed like 80 or 8080.
At the same time, in Azure WebApp settings, you need to define a variable 'WEBSITES_PORT' with the required value of the port
You can read it further here: things-you-should-know-web-apps-and-linux

Anonymous

2022-06-07T17:14:58.79+00:00

I've tried to add the variable 'WEBSITES_PORT' in the WebApp settings with the value 80, and I then specified the port 80 in my python script.
However, I still get the same error. (Maybe I misunderstood what you were suggesting and what was written in the page you linked though)
I compared my code to the one Azure uses as an example for the deployment of python (flask) webapp: https://learn.microsoft.com/en-us/azure/app-service/quickstart-python?tabs=flask%2Cwindows%2Cazure-portal%2Cterminal-bash%2Cvscode-deploy%2Cdeploy-instructions-azportal%2Cdeploy-instructions-zip-azcli

I did not find any meaningful difference between my app.py and theirs, nor in my html file, so I'm not sure if I'm doing something wrong there.

Manu Philip 16,986 Reputation points MVP

2022-06-07T19:21:50.497+00:00

I just have tried deploying a sample application based on Python 3.9 in Azure WebApp and it works perfect. You may try deploying the same app in my GitHub Repo
If it works for you too, I suggest to check the code in the application you are trying to deploy

Anonymous

2022-06-08T16:01:01.633+00:00

Sorry for the late reply, but I wanted to grasp better where the error might come from.
I checked your repo, and, coherently with what I already observed in the example flask webapp [here][1] https://learn.microsoft.com/en-us/azure/app-service/quickstart-python?tabs=flask%2Cwindows%2Cazure-portal%2Cterminal-bash%2Cvscode-deploy%2Cdeploy-instructions-azportal%2Cdeploy-instructions-zip-azcli I did not find any meaningful difference in the app.py and the html files.
I think however that the problem might be caused by the folder lib, in which all the core python scripts are included.
So the question arises naturally, can I upload custom libraries to a webapp? I hope you can do that, else I don't see how else I would need to rearrange the whole repo, it would be quite limiting.

Ryan Hill 26,146 Reputation points Microsoft Employee

2022-06-08T16:05:27.807+00:00

Can you post your application.py @Anonymous ? Your app is failing due a startup error:

2022-06-06T10:28:17.860765223Z /opt/python/3/bin/python: Error while finding module specification for 'application.py' (ModuleNotFoundError: __path__ attribute not found on 'application' while trying to find 'application.py'). Try using 'application' instead of 'application.py' as the module name.

Anonymous

2022-06-08T16:40:37.163+00:00

I cannot upload any file (I tried with txt, py and html) and the script exceeds 1600 characters to be uploaded in a single message. That's quite limiting really, if you have any suggestions better than uploading the script in multiple messages let me know:)
Sign in to comment
Anonymous

2022-06-10T12:03:15.223+00:00
Thanks a lot @Manu Philip & @Ryan Hill for helping me out.
After a bit of debugging, I managed to get the app to work.
The main problems were:

The app needed to have this tree structure (it is very important to have the __init__.py file inside application):
/app.py
/requirements.txt
/application
---/__init__.py (in which I just wrote 2 lines of code (from flask import Flask and app = Flask(__name__))
---/lib
---/static
---/templates

Open-CV was crashing the application in the Azure machine, so I needed to migrate some code to PIL&numpy instead of cv2.

The deployment was not successful using GitHub but for some strange reason it was using VIsual Studio Code and Azure tools.
Please sign in to rate this answer.

1 person found this answer helpful.
Ryan Hill 26,146 Reputation points Microsoft Employee

2022-06-11T02:21:41.587+00:00

Definitely seems like your issue is similar to https://stackoverflow.com/a/52317799/10987310. Glad you were able to resolve it. From your initial logs, your GitHub action was deploying correctly, but your application was failing to start because application wasn't a loaded module. I would say you can still utilize that pipeline.
Sign in to comment