I'm curious how a magnetic fluorescent ballast can get EOL.

Moths ago, I finded a 50~60 years old 40w preheat fixture, and it's fully working perfecly. The magnetic ballast have a limited time duration?

Can, and it is not that " difficult":
It could be iron corosion, causing it to expand and so overstress the winding and so damage the insulation.
It could be the impregnating material degradation, making the winding loose, so allow it to vibrate by tze magnetic forces, so rub away the insulation layer.
Or break the wire.
Or UV caused corrosion of the organicmaterial parts (the terminal support, the winding impregnation, causing their degradation and shorts).
Or the terminal corrosion causing bad connections.

It could also be caused by overheating that could be the result of operation at a too high ambient temperature or a long time on stuck starters.

How many years of use have a magnetic ballast? For example:
40w 220V AC 50Hz ballast with a F40T12 tube running in a 220V AC 50Hz power.

It's hard to tell exactly. There are many, many variables here. It depends really on all factors that Medved and I have mentioned. Such a ballast that would be operated in a relatively cool, low humidity environment and never left with stuck starters will likely last upwards of at least 200 000 hours if not more. A ballast that is used in a humid hot environment and left every now and then with stuck starters for a long time, might not last more than 80 000 hours. That being said, there are no hard and fast rules for estimating such lifetimes. Manufacturers test using certain temperature parameters and furnish approximate life based on that, but many installations differ quite a bit from that.

Generally design goals are (and always were) around 100k hours before the EOL effects start to appear in any significant number and then obviously the lowest cost.
The approach obviously is to design it as durable as possible if it does not mean any penny extra in cost. That for instance means the modern low loss magnetic have very high chance to last long in dry and dark (shielded by a cover from the lamp UV) environment (where the corrosion problems are eliminated), because the low loss requirement makes them to run rather cold (there the "10degC colder = double life" rule apply), but the materials are already the cheapest possible.
The electronic are other story - these have rather high efficiency and allowing some component to run hotter means saving on cost, so these will really start to fail after the designed service lifetime and it is unlikely these will last much longer without starting to make frequent troubles (given the depth of understanding about really many aging mechanisms so the ability to design the things "just enough" to meet the requirement with the lowest cost).

In real life the typical life limiting factor of magnetic component of the mains frequency ballasts use to be just only the corrosion and with some bad fixture designs the stuck starter damage and that use to be in the 10 years ballpark. And that is in most cases not because of the ballast itself, but the fixture degrading so it is not able to provide adequate protection to the ballast.
A way larger failure rate comes from capacitor failures, these tend to be way shorter lived and usually you have to replace 2 or 3 capacitors. It could be less, if the capacitor is not essential for the main ballasting function, so when you may easily tolerate 20% or more failed ones in the installation (usually the case for a PFC) without affecting any fixture operation.