NASA really needs to just fill their telescopes with gyros upon gyros, like just plaster the things with gyros and reaction wheels for 50x redundancy since that's always the mission critical thing that fails first. Or maybe a xenon powered ion reaction control system or something that would last longer then hydrazine.
It’s an interesting tradeoff. Missions are designed (and costed) for a nominal mission lifetime; adding more redundancy increases costs. But it’s true that the successful missions tend to stay in operations much longer than their normal lifetime.
There's a limit to how much redundancy you can put in a satellite, but with how consistently these parts fail it would make sense to put in a lot. If 1% of the Hubble's weight was extra gyroscopes it would have 30-40 of them.
But I'd expect it's often cheaper to send up a new telescope than to repair an old one.
Sending up Hubble II and Hubble III would of course not mean that Hubble I stops being used.