Since most posts here are about things that have gone wrong, I thought it might be interesting to post about something that went right...
We had a "moment" last week when our monitoring system reported the load on one of our mailbox servers was above 40 (and rising).
CPU utilization was about 20%, and (disk) wait % as reported by top was between 3% and 5%. The UI was just a hair slower, but this would not have been noticeable on a slow Internet connection.
We traced the cause to a customer doing a several GB pst import; the load was from Zimbra indexing the messages.
Nice to know that Zimbra handled this in a way that didn't impact the other end users on this mailbox server. Kudos!
All the best,