I've just noticed some server downtime, apparently only on webmail, with the following error on the logs:

MDB2 Error: unknown error (-1): _doConnect: [Error message: Too many connections] [Native code: 1040] [Native message: Too many connections]

I also spotted a few of these:

roundcube: PHP Fatal error:  Maximum execution time of 120 seconds exceeded in /var/www/roundcubemail-0.2-stable/program/lib/imap.inc on line xxx

After restarting mysql, it restored the webmail service. However I would like to investigate what could have caused the issue.

Looking at the root bash history I noticed that the SSL certs in /etc/pki/iRedMail/ were accidentaly moved (rather than copied) to another location by a new server admin. I believe this was around the time that the issue started. I've now moved the certs back in place, but I cannot think how moving the certs could impact the DB in any way.

Looking at the mysql log doesn't show any evidence of anything.

It is fairly safe to assume that we don't have enough concurrent users to cause any DB issues.

Any ideas as to what could have caused this, or hints on where to look further for clues would be greatly appreciated.