Yes we are trying to investigate what exactly is happening. Issue is it is on production and happens rarely, but when it does happen, we have thousands of internal users that lose access. We're just trying to get something in place for it to recover while we track down what the issue really is. Thanks for the info @gregrobson!
We are having an issue every once in a while in which the database connection goes down. We can catch the exception, and would ideally like to force it to reconnect. Currently, the only way we know how to fix it is to restart Apache.