My armchair speculative guess: there are only 4 bolts effectively holding the door onto the roller pins and lower hinges. Somehow I don't think it's a simple case of someone forgetting to tighten the bolts; since they're using castle nuts, forgetting to torque them down would leave behind extra cotter pins. Those bolts look positively tiny, probably no more than M12 diameter, and are subject to intense shearing forces. In the case of the upper roller pin with locking bolt, they are effectively two cylinders perpendicular and on top of each other, which causes extreme compressive forces to be concentrated on one tiny spot on the bolt. My guess is they cheaped out and switched to an inappropriate/softer bolt, which sheared and/or corroded. If one shears, the load quickly spreads to the remaining 3 bolts which all also shear off. This theory would still jive with OPs article about UA; "loose bolts" may not necessarily mean "untightened nuts and bad QA", but rather, signs that the locking bolts are all beginning to bend or shear.
Longitudinal forces on the bolts there, not shear. A change during construction doubled the force on the nuts and the bolts went through weak weld joints in beams. (Also not a PE.)
Wow
or if some other mechanism like the springs on hinges below grew weaker over time .. would that in turn put too much weight on the bolts?
what if the springs on the lower hinges (that are supposed to keep an upward pressure on the door, supporting its weight) are weaker than necessary or somehow failed and went unnoticed... would that put too much burden on the bolts and other fastening mechanisms and cause them to fail in turn over time?
On paper all 6 degrees of freedom may be constrained with the plug but in principle compliant pads may allow shear forces and/or rattling on the fasteners and they eventually fail.
I will eat my hat if the eventual required solution doesn't involve some kind of mating cup/cone or wedge shaped parts to more positively constrain the plug.