Because:
- We keep seeing random timeout errors in production.
- It's better to retry operations that time out a few times before
  giving up completely - this runs in the background sending cron mail,
  so we should try to take care of our own problemns ;-)