There should always be time to do some burn-in testing for a new production system. Even if it is only a couple of days. If you look any server life analysis with decent sample size you will find a good number of early failures followed by a couple of years with very low failure rates before the rate of age and wear related failures starts to rise again. The purpose of burn-in testing is to flush out those early failures so that they do not happen in production.
I'm sure you know that distributed computing apps are great for burn-in testing -- Got some spare cpu cycles ? Join Team Helix or Team Starfire!