We need to switch over to flask style app based logging because Since Flask 0.8, if exceptions are not set to propagate, unhandled exceptions are sent to the application error log.
A quick fix would be to set PROPAGATE_EXCEPTIONS to True in the Flask config as this will enable propagation of exceptions to the python logging library. To avoid future pain it makes sense to push the logging to the app itself.
This may push all logging to the logs in /var/log/supervisor/copilot.... by pushing it through gunicorn. I don't have a problem with that even though debugging through /var/log/copilot.log and supervisor has been easier because of the separation of the logs so far.
We need to switch over to flask style
appbased logging because Since Flask 0.8, if exceptions are not set to propagate, unhandled exceptions are sent to the application error log.A quick fix would be to set
PROPAGATE_EXCEPTIONSto True in the Flask config as this will enable propagation of exceptions to the python logging library. To avoid future pain it makes sense to push the logging to the app itself.This may push all logging to the logs in
/var/log/supervisor/copilot....by pushing it through gunicorn. I don't have a problem with that even though debugging through/var/log/copilot.logand supervisor has been easier because of the separation of the logs so far.