I’ve built a streamlit app deployed on an Azure VM that wrangles data pulled from various APIs and generates a standard stream response to the user. The app, deployed on a conda
environment, crashes after a few minutes to hours after deployment on my instance. I’ve tried to monitor disk usage using du
and it doesn’t seem to go over 10% of CPU or VRAM usage at any point in time. I’ve also tried to monitor logs using the --logger.level=debug
flag on streamlit run
; but it doesn’t seem to record the crash. Has anybody faced similar issues deploying outside the Streamlit cloud? I’d be eternally grateful for tips!