部署至Azure WebApp后无法导入Celery的问题排查求助

Hey, I’ve run into almost identical deployment headaches with Django + Celery on Azure WebApps before. Let’s break this down step by step to track down the root cause:

1. Verify Celery Version & Installation Integrity #

The most common culprit here is a version mismatch between your local venv and Azure’s antenv. Different Celery versions have different structures in celery/__init__.py, which can cause the Celery class import to fail even if the file exists.

• First, run pip freeze | grep celery locally to note your working version.
• Jump into Kudu for your Azure WebApp, run /home/site/wwwroot/antenv/bin/pip freeze | grep celery, and compare the version numbers.
• If they don’t match, explicitly pin the working version in your requirements.txt (e.g., celery==5.2.7) and redeploy.
• If the version is correct but the install seems corrupted, force a reinstall via Kudu:

  cd /home/site/wwwroot && antenv/bin/pip install -r requirements.txt --force-reinstall
  

2. Check for Naming Conflicts in Your Project #

Even though you renamed celery.py to celery_setup.py, leftover files or conflicting modules can still throw off imports:

• Use Kudu to search for any stray celery.py files in your project:

  find /home/site/wwwroot -name "celery.py"
  

• Delete or rename any unexpected matches (old deployment artifacts are a common offender here).
• Also, double-check that none of your Django apps have a celery submodule (e.g., an app directory with its own celery.py or celery/ folder).

3. Confirm Python Path Configuration #

Azure’s WebApp environment might not be adding your project root to sys.path correctly, which can lead to unexpected module resolution:

• Add temporary debug code at the top of your wsgi.py to log path details (deploy and check the application logs afterward):

  import sys
  import celery
  
  print(f"sys.path contents: {sys.path}")
  print(f"Celery module being loaded from: {celery.__file__}")
  

• If your project root isn’t in sys.path, manually add it in wsgi.py before loading Django:

  import os
  import sys
  
  # Add project root to Python path
  sys.path.append(os.path.dirname(os.path.abspath(__file__)))
  
  os.environ.setdefault('DJANGO_SETTINGS_MODULE', 'your_project.settings')
  # ... rest of your wsgi code ...
  

4. Validate the Azure WebApp Startup Command #

Sometimes Azure’s default startup command skips activating the virtual environment, causing the system Python (not your antenv) to be used:

• Head to the Azure Portal, navigate to your WebApp’s Configuration > General settings, and check the "Startup command". It should look something like this:

  source antenv/bin/activate && gunicorn --bind=0.0.0.0 --timeout 600 your_project.wsgi:application
  

• Make sure the command explicitly activates the antenv before starting the WSGI server.

5. Try Downgrading Celery (Last Resort) #

Python 3.7 is end-of-life, and newer Celery versions (5.x+) have dropped full support for it. If all else fails, try pinning to a 4.x version of Celery that’s compatible with Python 3.7:

• Update your requirements.txt to include celery==4.4.7, redeploy, and test.

内容的提问来源于stack exchange，提问作者Ross Halliday