Did a lot of reliability testing during my working career. At one place I worked, I designed hip implants and tested them on an Instron testing machine. We would apply a load of about 500 lb, cyclically, at a frequency of 10 Cycles/ sec to the top end of the implant that was cemented into a fixture. These implants were all made from the same forging and then machined to the exact same specification on the same machine. We tested them until they broke due to fatigue. The number of cycles it took to break the implants varied considerably. The results followed your typical bell curve. With the mean around around 13,000,000 cycles. But we did have one that broke at below 10,000,000 cycles and a few that never broke (went up to 20,000,000 cycles. The molecular structure of a metal part consists of many slip planes. The molecules are not all congruent and perfectly aligned. These discrepancies cause the working life to vary considerably. Engineers design parts to have a typical life for a particular application. The mean time before failure is usually a well acceptable time frame. Even the best designs still have alot of variability. You will always have parts that last forever and parts that break early on. It is usually not a problem, unless you start to see alot of failures that fall outside the typical bell curve.
The failure seen here is not a common one. Hard to say from the info if its a defect of the metal or something else was going on. An investigation into it should easily illuminate the exact cause. It sucks when its your bike. I bought my first Ducati, a used monster with 4,000 miles, years ago. 1 week after I bought it, I changed the oil and found the left main bearing retainer in pieces in oil the pan! Not good.
