It's actually not creep. It's a phenomenon called stress relaxation. Creep is when strain occurs in a material under constant stress/load. Relaxation is a change in stress/force while maintaining a constant strain. In the case of the spring you strain it to whatever compressed length and it sits there a long time. During this time the spring essentially suffers from some plastic deformation while the stress is under the tensile stress of the material. This is why creep and relaxation are often confused. They are related phenomena, but different a bit. Additionally when you move a spring back and forth it also undergoes fatigue. Certain alloys are more or less susceptible to fatigue. Additionally, temperature, number of cycles, etc all play a role in how quickly both relaxation and fatigue affect a metal.
Generally what caused failure in a moving part is fatigue. Dislocations in the crystal structure form after many cycles within the elastic range. This leads to an increase in brittleness. Theoretically one could re heat treat the part after say 5000 rounds and this would greatly increase the life of the extractor by relieving any internal stress and dislocations formed in the crystal structure due to fatigue. Realistically, someone should just start to manufacture new quality extractors and make a hell of a lot of money.
Sorry to nerd out on ya. Just in case you or anyone else was curious.
I'm a structural engineer with a master's degree in civil engineering.
Let me think on this for a tad.
Here we have a spring (UZI extractor) that's left in its little hidey hole in the bolt for a decade or two.
If it exerts less sideways force on the 9mm case than it did originally, that would be stress relaxation.
If you take it out and it has a different shape than it originally did, it's less bent, that's creep.
Can you have stress relaxation in a spring without creep? What would be an example of that?
If you leave a gun magazine loaded for 20 years and you take the spring out, the spring will definitely be shorter than it was when it was put in the mag. With the creep accounting for the lower force exerted on the follower.