Azure App Service is a service used to create and deploy scalable, mission-critical web apps.
Thanks for the troubleshooting suggestions! I confirm the following:
- requirements.txt at deployment zip root
- SCM_DO_BUILD_DURING_DEPLOYMENT set to 1
- Python 3.11.14 runtime working correctly
- Startup command: python -m uvicorn research_agent_ai_assist:app --host 0.0.0.0 --port 8000
- Log stream shows successful Oryx build start
As a test, when I reduce to <10 essential dependencies in requirements.txt, uvicorn gets installed
Collecting uvicorn... Successfully installed h11-0.16.0 uvicorn-0.38.0
If I load the full 68+ dependencies (LangChain, pandas, matplotlib, scikit-learn, chromadb, spacy, etc.) in requirements, I think the Oryx build process times out during pip install on Free Tier but I am not sure...
Is my Free Tier F1 simply insufficient for scientific/ML apps with large dependency sets, or are there strategies to handle 60+ packages within the resource limits?
Thanks