For safety critical systems, one strategy is to store at least two copies of important data and compare them regularly. If they don't match, you either try to recover somehow or go into a safe state, depending on the context.
Yes, but what's easier depends on layout. "Consensus" makes me think of multiple entire nodes, and in that situation you can have a nice symmetry by making each node store one copy and one small hash.
If you're doing something that's more centralized then one hash might be simpler, but if you're centralized then you should probably use your own error correction codes instead of having multiple copies.
Seems like chronometers would be a case where two are better than one, because the mistakes are analog. If they don't exactly agree, just take the average. You'll have more error than if you were lucky enough to take the better chronometer, but less than if you had taken only the worse one. Minimizing the worst case is probably the best way to stay off the rocks.