how did you check it? just to point out that if you have meter leads in play with the measurements or even the ampmeters own resistance, you have added some resistance. so when you REMOVE the meters resistance, it goes up even higher than that.
if you have a bench power supply you can put in a simulated voltage into the "head" of the light and get a reading off the current from the power supply, then the lead and meter resistance doesnt mess up your numbers as much.
how long it takes to die is relative to the cooling it gets, how long it stays at the high voltage, how well it is attached to the sinc. but yes it WILL be being hurt, and when it gets hurt, it will reduce a bit in output too, as the tiny molecular gates fry they dont any longer put out light.
adding in the smallest ammount of resistance like say .25-1 Ohms would solution the Major problem, still give you excesses and probably fix it up fine. or you could use one of them 8x1735amc chip boards, and it would clip the top end in similarity, and there would be less issues with guessing the actual amperage with that too.