Fenestrons are great but pilots have to understand that they behave slightly differently from a conventional tail rotor.
Some of the anti-torque thrust comes from what has been described as 'lip-lift' where the flow around the duct adds to what is being produced by the fenestron itself - a bit like a coanda effect.
When you get a situation where something affects that flow and disrupts the 'lip-lift', the demand for power pedal is suddenly increased to compensate and, if you are not on top of your game (or ahead of the aircraft) you can end up with a left yaw you weren't expecting.
There is still more than enough TR authority but it requires you to use a bit more pedal - this can sometimes be masked by the TR linear actuator if you have SAS or AP in yaw.
This is where the myth of Fenestron Stall was born - and subsequently disproved by Aerospatiale TPs. It hasn't stalled and you don't have a TR malfunction, you just need to apply more right pedal - perhaps all the way to the stop!
This isn't usually a problem but at high power, many pilots are reluctant to use that extra pedal due to the marked Tq spikes (also typical of fenestron equipped aircraft) experienced if you are less than gentle with the yaw input.
You can see this effect when you accelerate sideways (or hover crosswind with the wind from the right) - eventually you get to a point where the airflow through the fenestron needs to reverse as you apply lots of left pedal - there is a marked yaw disturbance but you still have full control and as you pass through this, normal feel is resumed. The same happens as you slow down again and that is where the Tq spike can occur, as you push right pedal to re-establish 'normal' flow again.
Airbus put out a safety notice a while back emphasising that the thrust from the fenestron vs pedal position wasn't as linear as from a conventional TR but slightly S shaped - for the reasons mentioned above.
If the pilot in this case, in gusty winds and manoeuvring to land, experienced what he thought was a TR problem, it might go some way to explaining the result.