The trick to resilience, and what makes it an important concept, is that we don't need to know the specifics of a crisis in order to know some general things about what will help us deal with it. By looking at what things have been helpful for withstanding past crises (including ones that were major surprises at the time), we can deduce generalized crisis-handling capacities.
For example, one such generalized crisis-handling capacity is resource buffers. A crisis is likely to demand additional resource expenditures to handle, or even to directly attack and reduce the resource base itself -- whereas the reverse is highly unlikely. So if a system limits its resource use to something less than what would be optimal in a crisis-free world, it will, ceteris paribus, be able to weather the crisis better than if it had been straining its resource base to the max.
Democratic information processing is another generalized crisis-handling capacity. It would be easier to handle any crisis -- whatever its nature -- if the system gets an early warning and full information, which we know from past experience is more likely to happen when hierarchies don't restrict the flow and sharing of information.
Diversity -- genetic, cultural, psychological, etc. -- is another useful generalized crisis-handling capacity. An un-diversified system may be optimized for the pre-crisis conditions, but a crisis necessarily changes those conditions. If the system is diverse, there is a greater likelihood that the answer to the crisis is somewhere to be found within the system already
Resilience is always a matter of degree -- no system is perfectly resilient to every possible crisis (though Haab seems to think such a thing is being demanded), and increases in resilience often come with costs (e.g. in the form of foregone profits from leaving a resource buffer). That leaves us with an eminently political question of how much of various types of resilience we want to build into our social system.