When the bot listens for a long time, it will occasionally fail silently or with non-exception-throwing errors. This is due to Facebook either flagging the account or simply invalidating the session (causing the connection to become stale).
The deployment process can handle exception-throwing errors by automatically restarting the bot, but it cannot detect these types of silent errors. The monitoring module attempts to handle this, but unfortunately it is not very effective in doing so.
Look into ways to improve the monitoring module to detect and handle these errors to allow bots to run for long periods of time without manual intervention.
When the bot listens for a long time, it will occasionally fail silently or with non-exception-throwing errors. This is due to Facebook either flagging the account or simply invalidating the session (causing the connection to become stale).
The deployment process can handle exception-throwing errors by automatically restarting the bot, but it cannot detect these types of silent errors. The monitoring module attempts to handle this, but unfortunately it is not very effective in doing so.
Look into ways to improve the monitoring module to detect and handle these errors to allow bots to run for long periods of time without manual intervention.