Look at the linear graph. It’s probably flat. I’m willing to bet that with that 3rd peak below -50db it won’t show up or be significant enough to waste the processing power on.
Just for giggles, look at the predicted output and take out the second harmonic.
Play with it and maybe try things like setting the 1st and 3rd, and leave out the second. The predictive graphs are really accurate and if anything a touch conservative IMHO.
Most assuredly not. You want the minimum notch footprint to get the job done.
Here is another experiment you can do. Disable the notch filter completely and look at the estimated post-filter result from the Gyro filter only. This will give you some idea of what really needs to be configured for the notch filter.