Because you have windshear and you have windshear.
Not every windshear is a microburst, which is what we practise in the simulator.
No microburst in the EK accident.
If I divert every time the wind shifts/varies by 10-15 kts, I would not be doing much else.
I think this a very pertinent observation. The wind effects around DXB at the time were most likely due to a slow-moving sea breeze approaching the airport from the NW. Shear associated with a convergence will be positive (performance enhancing) whichever way you go through it as opposed to a microburst which is a divergent phenomenon.
The initial symptoms will be the same for a convergence as a mature microburst, i.e. gain of airspeed, tendency to go high, unusual thrust lever position, etc. so people who have been trained that windshear = microburst will initiate a WS GA. As the approach may be unstable by then, it’s not a bad response and errs on the side of safety.
I have to say I am somewhat skeptical as to the reports that say the aircraft sank back onto the runway due windshear alone. It would have needed to be pretty extreme to do that to a 777 below MLW with firewall power on both engines, even given the high OAT. Consider that at that weight and temperature it should be able to climb with reduced thrust
on one engine after the other one failed at V1. A shortfall in thrust for part or all of the manoeuvre is a simpler explanation, although the atmosphere can throw some right curve balls at you on occasions...