Comment 6 for bug 1549516

there is another potential way to mitigate this more short term, which is to do the "SELECT 1" ping only some of the time. E.g. if a pool checkout occurs and we see that we've been "idle" for some time, we do a SELECT 1, then for subsequent checkouts within a short time period from that one, we don't do it again. This would cover the case of the MySQL database being restarted and allowing the app to recover without error once the DB is running again.