I read somewhere that they were not sure yet about all the failure modes This is in the rhelm of lawyers and big money so they are not motivated to say more than the minimum.
All relays will fail eventually. The design criteria certainly includes a minimum life cycle. So many hours of operation or so many cycles before a failure. They do testing on the early production units to ensure that it meets the requirements. And because its a life/safety item, I am sure that they have designed into the relay a failure mode that does not include catching fire. It should be designed to stop passing current at end of life. Instead, apparently it has an unplanned / unknown failure closed mode that allows it to continue to pass current upon failure. This should not happen.
The criteria they are using for the unplanned, catastrophic failure mode is based on examination of the design, and of failed units. They have done the modeling and concluded that if the relay has operated correctly for 600 hours minimum then it is not going to have the fail closed mode. It may fail, but it will fail open as it is designed.
Also, the heat cycle test probably has two failure modes. One is if the temperature exceeds a set value during any of the ten cycles. The other one is the slope of the delta T as the temp increases across the ten cycles. If the slope of the curve (rate of temp increase from one cycle to the next) exceeds a threshold then the relay is faulty.
The decisions around the recall are made by executives but the criteria for the testing is made by engineers. The engineers want the owners to be safe and secure. they want the truck to perform as it was designed.
Give them some credit