The concept of an open beta stress test isn't an unusual one for Blizzard - it does pretty much what it says: Tests the servers to find out if the game can handle the amount of traffic it will be put under at launch. The theory says that the number of gamers logging in during a brief period is pretty similar to the number who'll be logging in on Day 1, so it provides an accurate snapshot of what will be happening.
In some ways, this test is even more worrying given Blizzard's history with stress tests. The World of Warcraft open beta was held back in 2004, with many gamers reporting no complaints, long wait times or other drama. When the game launched, however, nobody predicted its popularity, which saw World of Warcraft crippled under the strain. Within a month, Blizzard was prompted to double the number of WoW servers, but it took the company six months before it could guarantee a stable online experience.
With Diablo III, we're potentially looking at a similar outcome, due to the seemingly opposing test results. If the server cannot handle one weekend's worth of rabid fanboy traffic, can it really be anticipated to look after the inevitable deluge when the game launches?