At the debrief the engineers bounced around; aerodynamics and structures with expressions of surprise and relief that everything held together!
That's rather concerning given that they should have the best understanding of the safety margins built into Vd, which are considerably greater than .005M. Whilst we should always aim not to eat into safety margins, that is what they're there for.
My experience - admittedly mostly part 23 and smaller, is that we do fly towards Vd in prototype testing. Typically if we get to Vd, we do not pass, and we declare Vdf as co-incident with Vd. Not uncommonly, we hit something of concern (I've seen flutter, canopy distortion, aileron reversal, longstab reversal...) below Vd - at which point we stop and analyse. If the value of Vdf is acceptable for the desired in-service envelope, we have usually stopped at that point, whilst if it isn't, we go looking for solutions.
Of course then, in defining Vne as a margin (usually 10%) below Vdf, we then create a further safety margin. Most regimes will then only go beyond Vne ever again under formal flight test conditions and all in-service flying, including routine air tests don't exceed Vne - that said it's understood that the Vne to Vdf margin is a safety margin, so small Vne exceedences aren't (usually) a cause for concern.
Swap "V" for "M" where required.
G