Improve App Robustness by Handling and Logging Panics #62
Description
Currently, if the app crashes due to a panic, which it has done on occassion for unexplained reasons, the panic info we would need to debug it is not written to the log. Also, it is not automatically handled or restarted, so the app just dies.
Two steps should be taken to improve this situation.
- redirect the panic output from stderr to it's own log file, so it is available in the event of a crash
- Use systemd, docker, or some other management type service to automatically restart the app in the case of a crash.