I’m not positive about the construction, but I think the centering alignment is probably important. You likely need some kind of 2 part specialty epoxy.

In my experience as a maker with lots of experience playing with finishes and epoxies (owned an auto body shop), this is a rabbit hole that will cost more than just replacing the driver.

When it comes to adhesives like this, cleaning and the proper tooth for the adhesive to grab onto, like the proper sandpaper grit finish, are critical and the primary factor. However a very close secondary factor is choosing an adhesive that bridges the expansion rates of the two joining materials with enough ductility to compensate for the differences. If one of the materials expands a whole lot more than the other across the temperature range the object experiences, the adhesive must compensate for this stress or fail. Most super glues are formulated to cure aggressively fast. This makes them brittle, or in other words they have very little ductility. I expect the ceramic of the magnet to have a very different expansion rate to the metal of the driver. I suspect this failed from a combination of bad initial prep of the metal driver base combined with temperature over time. I don’t think you are likely to have success with a random super glue with unknown properties.

I am no expert in the engineering of audio drivers, but I hope this info is peripherally helpful. It is accurate to the best of my understanding.

Try super-gluing it! You can see how to align it by the hole/mark. Not that it should matter.

Make sure the surfaces are cleaned with isopropyl alcohol first, and use a think even layer of glue (a razor blade can smooth it out).

Thinking about this a bit more… assuming the driver still technically works, I think the driver manufacturer was just trying to glue on a second coil magnet to couple the flux with the actual coil magnet. While the actual coil magnet is located with precision I don’t think the one on the outside is super critical.

My ability to diagnose a finish/adhesive failure over an image on the internet is extremely subjective and limited. I can tell the failure was not likely to have been caused by the adhesive because of how it transferred onto the magnet. The driver side looks very suspect; like a surface that has had insufficient cleaning and prep. Professional finishes level cleaning prep is extremely rigorous. It is not quite chemistry lab/clean room levels, but it is unlike anything in a typical domestic environment.

It doesn’t hurt to try this if you have the supplies: Clean both surfaces well with acetone or lacquer thinner, with the latter being the better choice here. Then use a sandpaper between ~180-600 grit, with 400 grit as my likely go-to, and take off all of the epoxy from both surfaces completely. Both sides should be a consistent matte appearance. Then I would use whatever 2 part epoxy I and find that is intended for metal, might be thinner in consistency, and has a long cure time. Use a piece of cardboard wrapped around the driver for alignment, get a consistent coating of epoxy across the entire surface and give it double the specified cure time or even more if the ambient temperature is low and/or if humidity is high.