This is an old chestnut. When you discount aircraft that are undergoing servicing, waiting to undergo servicing (and cannot be flown in the meantime due to hours in the log) and other aircraft set aside from normal flying due to trial and test equipment on board, you are not left with many.
The only stat worth pursuing or monitoring, in any fleet of airframes, is the number (percentage) of aircraft that are not available due to rectification. Even then you need to be more specific about rectification. As I write, I am quite certain that a jet is having a black box (associated with its mission, not its airworthiness) replaced, or waiting for the replacement. While this routine repair is going on, the aircraft is not available because of the open entry in the maintenance log. Most of these routine faults do not have any real effect on the aircraft's availability to fly. Clearly there will also be other faults that are unsafe to take airborne, but most of them are repairable within hours (tyre worn outside limits, fuel coupling leak, generator fault) and while they make the aircraft unavailable while they are being fixed, they are not having any effect on the flying programme because the aircraft is also receiving routine servicing simultaneously. Its common throughout the world as a front line system.
Where it becomes serious is when the defects cannot be fixed within a day (or so) and the aircraft has to come off the front line and wait for spares, manpower or space in a hangar, etc. Its not that the aircraft is necessarily in a dangerous condition; it just need some increasingly scarce resources. Today, at a guess, I think there is less than 30% of the whole fleet that meets this narrow category, and its the only category that really matters to the planners because we cannot plan for it, airframe by airframe. Add to that category the previously mentioned reasons for unavailablity and we get to the two thirds unavailability.