While I like the NOTAR system a lot, the 902 anti-torque is less immediately responsive than most conventional tail rotor systems I've flown. When at MAUW and hot, even at lower altitudes you can find yourself applying a surprising amount of left pedal when flying helipad profiles or HOGE with a wind from the right, or arresting a clockwise yaw. It really wouldn't surprise me at all if at 11'000 ft PA on a +ISA day you might easily run out of left pedal. When you pass around three quarters left pedal there is a noticeable audible change in the aircraft as the NOTAR fan blades' pitch increase towards max - something that personally I think is apparent in this video whilst the aircrwaft is yawing to the right. My feel is that he's sat at full left pedal in efforts to stop the yaw, but it's not working.
Crab, generally no problem with NOTAR at low power settings but I have no first hand experience of the impact of a strong/gusty updraught on the effect of coanda in the hover. It's an interesting question - the coanda effect alone is providing approximately 60-70% of the anti-torque in the hover with just 30-40% provided by the thruster can (or so they say). If the updraugh were to significantly disrupt the required flow around the tailboom, enough to reduce the coanda, then it would put significantly more demand on the thruster drum which was probably already at full chat. That said I still think it was more likely a case of WAT and running out of pedal.