It depends on the context. Personally, the trigger points are: - ever evolving problem domain research manifesting as architectural doubt - underspecification

It depends on the context. Personally, the trigger points are: - ever evolving problem domain research manifesting as architectural doubt - underspecification - unforeseen integration issues with host API or environment (pattern collision, resource collision) - F.O.M.O i.e. ego corrupted view of the problem - cognitive fatigue I’m sure the list is as endless and diverse as the people who make software...