Failure reporting and random shutdowns #33
Description
The production instance's mysql server has crashed twice in the last 6 months, taking down the website unexpectedly and silently. The issue is easy to identify, if you can visit the timeline at baly.kenyon.edu/timeline but not the homepage, then this is very likely the issue, since the timeline (and about) pages are static and do not request from the database.
For solutions, we could try updating the mysql version, although any dependency upgrades will be difficult with the limited testing at the moment. A more immediate step would be implementing error reporting to email a maintainer when a server error is encountered (which could help with bug-fixing overall), and eventually pinging the sql server periodically (rescuing? Would probably need sudo though), catching these problems before anyone is able to see them.