I guess that question is not technological feasibility, but financial feasibility and applicability. A single satellite lasts a long time in orbit. By the time its fuel is gone, the satellite technology has changed so dramatically that the old satellite would have become so obsolete that it becomes meaningless to play with the old one.
Much like the HDTV's nowadays. The technology changes so fast. By the time your tv is broken, much better tv has come out. No one wants to fix their broken tv anymore. its much easier to simply get a new one. No one wants to be stuck with an old tube TV anymore.
No one (in the western world and maybe China) bothers to repair TVs because TVs have fallen so much in price. I doubt many people would take such a wasteful attitude with more expensive assets cars for example.
Satellites and space vehicles are by no means cheap, at millions of dollars a pop. Average life for LEO satellites is 5 years, and 8 for GSO. With such short lifespans, nor do obsoleteness affect them much.
Few countries' militaries still possess sub metre resolution optical imagining satellites. Most countries are still launching brand new imagining birds with worse specs.
If you could top up your obsolete military satellites, you can easily transfer them to civilian agency (to save them having to send up a new one with the same or worse specs) or even sell it on and potentially make a huge chunk of the original build and launch costs back.
If you take all those savings and invest in modular satellite design and an orbital satellite repair, upgrading and recycling station.
Yes, it will cost a lot to set up and run initially, but over the medium to long run, it should yield massive savings.
Applying such a measure to GSO satellites would also massively help to resolve the growing space junk problem building in that region because of the current practice of dumping dying satellites into 'orbital graveyards', since GSO birds are too far away to be pulled into the atmosphere to burn up like LEO birds.
