FAA and BOEING both knew a thermal runaway would happen
The standard set by the FAA for this safety critical event is one occurrence in one Billion flight hours. IOW not even once in the life of the fleet.
Boeing negotiated a lower threshold, one in ten million flgiht hours. A reduction in standard of 100:1.
There have been two occurrences in less than one hundred thousand Flight Hours. A further reduction in threshold of 100:1.
A very interesting read along those lines dealing with system safety asessments and how they failed is this great NASA report :
Why System Safety Professionals Should Read Accident Reports
The potential consequences are often recognized, but their likelihood is thought to be quite a bit lower than it turns out to be.
And by the way, the "standard set by the FAA" of once in one billion flight hours is for
catastrophic events, the one for
hazardous events is once in ten million hours. So far the events have just been hazardous. Yes, an emergency landing did happen, and some passengers were injured. But there was neither the loss of an airframe nor did any fatalities happen, so formally once in ten million hours is the "correct" number to use. If only that number would have been met...
So maybe here as well the aircraft was desgned in a way that the effects of a thermal runaway were correctly recognized, but the likelihood was underestimated.