@Reynard - is this vaguely your area ? design limitations of metals ? calculating safety factors
I break things, not make things, but yes, you would expect a reasonable Factor of Safety (FoS) to have been applied. At the Product Design Specification stage (PDS), the exact numbers for that particular geometry will have been worked out. The usual assumption at the PDS stage is that the end user of a product (not necessarily bicycle brake disks) will eschew maintenance and try and do daft things, and so a certain amount of "idiot-proofing" is built in.
Basically, the FoS is based on the load applied to a component, and how much stronger you want it to be than the maximum allowable load. However, as the component wears, the FoS reduces, simply because the maximum load it can take without failure also reduces. Modulus remains the same, but you've simply got less material taking that same load every time, which makes it more likely to deform and fail. And as you can see from the photo in the very first post, the braking surface has sheared off the support arms where they meet, as the force applied under braking was greater than what the worn disk could take.
That's simply been worn well past the time to change it.
Now I only have one bike with discs (mechanical, not hydraulic), and I don't do a particularly vast mileage, but I suspect that this failure is more likely to be down to persistent fouling or grinding away of material rather than purely high mileage. Expected mileage / lifespan / durability for average or typical use will have been covered at the PDS stage, and the spec should be more than adequate. Of course, the discs are consumables even if they are a component that lasts a reasonably long time, but just like with a car, any mud or grit caught in the calipers will act as a grinding paste, promoting excessive and premature wear. The other possibility is using pads that are too abrasive. Either way, the net result is the same.