I would prefer then that the challenge came in the form of a cave-in or a crop blight. Because if the challenge is predictable and constant, all it means is a bit more hassle in setting up a new system to deal with it -after which the game is no more difficult than before.
Actually, "crop blight" (or rather, pests in general, which can include things like locusts or fungal rot or just an infestation of caterpillars) is a major part of the continuing Farming challenge because of this.
The thing is, however, if you manage it well, resource limitations
should be a concern throughout all of your fortress's lifespan.
In games like, say, Supreme Commander, you have really only two resources, mass and energy, and energy is created through reactors you set up with mass, so mass is the real core resource. It's generated constantly, but the constant flow you get in is limited. Everything you try to build takes mass, but unlike most strategy games, you don't pay an up-front cost, it just takes, say, 10,000 mass total to build a thing, and a given worker can build 3 mass per second of materials, and you have an income of 120 mass per second. Once you have spent the total mass you need, you get your thing you were building. Dumping extra workers increases the rate you can spend your mass, but if you hit the maximum rate of mass income, it just slows all projects down equally.
As an RTS, of course, you're constantly building massive armies and throwing them at the enemy in massive hoards, so you need constant replacement units.
This works as well for the organic resources of farming, as food and wood and clothing and maybe organic alchemical resources if we get them in later will all be resources we are constantly consuming and replenishing, as well. That is the other point of the farming improvements; making fertilizers into a functional cap on the output-per-year of a farm.
This can be done to a degree with inorganic/mineral resources gained from mining. While technically not inorganic, coal, for example, is a resource you're going to want to continuously develop and consume. If weapons and armor degrade or need repair, metal may eventually be something that you need to have in a steady stream of supply, as well.
With all that said, however, I do believe that the notion of an unpredictable mining feature would be the best feature.
From a gameplay perspective,
we could start with geologic formations as well as 3d veins to make the actual topography of a stone layer a little more interesting. If we have unusual structures, it will help make the exploration of the otherwise generally boring stone formations more interesting, especially if we hit surprise magma tubes or high-stress formations.
We can augment these things with more deadly surprises underground, whether they are pockets of dangerous gasses, weaker stone that are more likely to collapse, tapping into a pressurized magma chamber that will explode if you mine too thin a wall next to one, or other things to keep miners on their toes.
Rubble, I still believe, makes the whole thing slower, which in turn, means that you will try to stop and make the choice of mining those things that give you the most return for your time, rather than just strip-mining everything without a thought, and hence, is required to make the rest of this more interesting.
Gasses and ventilation, likewise, is a long-term dream of mine, which adds many of the same logistics challenges that minecarts will add, although we will probably need either some sort of pumping system, or else a oxygen-producing plant that must be farmed underground to make the system have more meaning than just digging some extra air tiles.
As I said in the other thread, something I'd hopefully like to see mining eventually become is something more akin to a board game where you would have a choice of drawing face-down cards from different decks that promise different things, or perhaps more like a game of minesweeper, where you have clues in the rock that tell you if you are getting close to certain types of minerals or perhaps dangerous events, and you have to make choices as to what you're going to risk mining into and for what potential goods.