Maybe its the glue or gluing that is the issue. I doubt that you are overheating that motor.
I have actually glued in the magnets of some number of motor kits that I have built. It actually is a fussy enterprise.
First off the magnets either want to jump together, or push each other apart, so the only way to get the magnet inside the shell is to slide it in. I use a spacer---and I am sure the manufacturers are doing the same, probably with a dual fixture that also aligns the magnet lengthwise in the shell. I first put in the magnets of the same polarity --so in the case of a Scorpion, that would be 7 magnets, let say with the north pole pointing inward the shell. As long as the spacing is pretty uniform, there is no real difficulty there. Next step is to come in with the other 7. As long as you slide them in centered between the two adjacent magnets, it will easily go in. The spacer is a real help here.
Now the question is when do you do the gluing? If you use a low-viscosity thin glue (like thin CA), you can use capillary action to draw the glue under the magnets, especially if they are curved magnets that match the shell curvature. If you have flat magnets, then this wont work and you need to use something else. I used a rubberized loctite on one the Scorpion motor that I built for the Nobler (3020). However it really was too thick for capillary action to pull the glue under the curved magnets. That was the one that had a magnet come loose on me, but fortunately by a minor miracle I accidentally discovered it when I pulled the cowl off the plane for some other reason. When I glued the magnet back in, I also squeezed in the adhesive to partially fill the gaps between magnets.
So that is part of the issue. You are gluing two metallic surfaces together, not ideal I think for a CA like adhesive. It is important I think the make sure the surfaces are really clean.
On my last kit build, I used JB Weld. That also is tricky due to the thickness of the JB weld, and the way you need to put the magnets in. What I did in that case was to spread a very thin layer of JBWeld on the inside of the iron shell. I put that in my drill press and spun it, while also just warming up the iron shell with a heat gun. Not too hot, but just enough to allow the JB Weld to run a bit.
Then I put in the magnets as described above, but I also spread the bottom of each with another thin layer of JBWeld, trying to put a little more on the front of the magnet, since when they slide in the glue on the shell (and magnet) tends to get scraped off. It's a messy process. You want to keep any glue off the inner surface of the magnet due to the close tolerance of it to the stator head. Anyway after it was done, I took it all back to the drill press and spun it up. Again I lightly applied heat (didn't want to demagnetize those expensive magnets!) to allow the glue to flow. I actually spun it from both ends on the motor shaft so the the glue wouldn't puddle on one end due to gravity during the spinning process.
I think it turned out pretty ok.
I say all this to at least make a point that it is easy to screw up the process. I don't know what the manufacturers do, but I have read many stories of people having magnets coming loose (even on the stock Scorpions).
I don't think most of us are overheating the motors (except the Adamisons in their torture tests!). In my case the magnet came loose relatively soon in the motor use. I think it was broken loose by pulsing forces that the magnets must feel due to the commutation (14x the rpm). That's just a guess on my part. I am not sure if I have heard of magnets coming loose as a function of an aging effect, but don't know.