The original reasoning behind the 20000 ft limitation for flap extension (for Boeing at least) was that there was such a small probability that flaps would ever be required above that level, there was no point in extending the test program to validate handling. Perhaps Embraer thought the same. Certainly, had their aircraft exhibited any undesirable handling at FL200 during the test program, they would have either addressed it, or imposed a lower flap extension limit to provide a safety buffer.
Built-in asymmetry protection is expected to safeguard handling if the flap system goes out to lunch, but can't be expected to cover cases of other structural failure. One such case occurred with an old B732 which had the L.E. slat on one side partially break loose and wrap itself back over the wing during a missed approach. The crew found themselves in the classic 'between a brick and a hard place' situation. They could not land at the destination because the weather and runway length etc made that far too risky. But as they climbed out to divert to the alternate the aircraft quickly reached a speed where roll control was no longer possible. Fortunately, by very careful handling of speed, altitude and thrust they did make it to the alternate, but ate into their final fixed fuel reserve to do so.
A crew fixated on observing all limitations may well have decided to ditch. In the dark, over open sea, that would have been a far worse outcome.