ChuChu
The examples of cognitive overload discussed seem to focus on someone who's focused on a specific task they believe is going normally, to the exclusion of an input that would have warned them it wasn't.
I wonder how well the theory applies to someone who realizes something's wrong but doesn't know what it is. It seems to me that when you go into troubleshooting mode, as the AF 447 pilots apparently did, you're more likely to step back and look/listen for clues that might explain what's wrong.
The effect that is repeatedly observed is that the human finding something difficult starts to enter 'attentional tunneling'. This may be called several things 'the pilot's scan broke down', he had 'tunnel vision' etc.
The pilot with cognitive tunneling may concentrate on entirely the incorrect stimulus - it is not something that lends itself to logical decomposition.
Ideally, simulation rides should be set up to initiate attentional tunneling then the ride frozen or rerun in playback and the tunneling pointed out to the 'subject'. It was found that using this approach with military pilots reduced susceptibility to tunneling.
Yet again it is back to training.