I would not “pace” that to be nice - bad choices are rewarded with dire consequences.
Fair enough, but then you're still going to have your ogre problem. Because unless the PCs
really have to get to the Dungeon of Doom in a hurry, adding a 1 hour rest to their day or more of travel is going to make the resource expenditure against the ogres of little or no significance. (Maybe the fighter burns a surge during the short rest.)
Maybe somebody have already tried something similar and have some important learnings they can share?
Lots of 4e GMs have varied rest durations - eg short rest = 1 day, extended rest = 1 week, or = a few days in a haven (ie not just anywhere with room to lay down a bedroll).
And, as you've seen in this thread, some 4e GMs use encounters with fewer opponents and/or minions to introduce short or "filler" combats between the bigger deals.
And as I mentioned, 13th Age uses a "swap an early rest for a campaign loss" system that could easily be adapted to short as well as long rests.
But what I'm trying to convey in my posts is that the issue you're asking about - how to vary the ratio of combats to recovery - has a clear mechanical meaning in 4e: it's about spacing out the elements of an
encounter (in the technical sense of a resource-consuming event that occurs between short rests). What you are asking for is advice on how to have encounters that come in waves rather than all at once (ie there are discrete moments of combat, although in mechanical terms they're all part of one encounter). For instance, wo ogres on the way to the Dungeon of Doom, who suck an encounter power and some hit points that don't get recovered before the PCs enter said Dungeon, are - in mechanical terms - just one wave of the encounter that is (in mechanical terms) ongoing until the PCs take their short rest.
If you want to do this simply in terms of dungeon design, you build your whole dungeon using an appropriate encounter budget (say, Level +4 for heroic tier PCs) and then deny the PCs a short rest while they're in the dungeon. Whether you achieve this denial by way of fiat, or a 13th Age approach, or via random encounter tables, or via your extrapolation from the "living, breathing" dungeon environment, is secondary: choose whichever one works for you. The key thing, in order to get what you want, is the denial.
Now if you want to extend it to the two ogres, though, you're going to have to look more closely at how you achieve the denial. 13th Age-style will work. So will fiat. "Living, breathing" adjudication might - but equally your players might find it a bit odd that they can
never get a peaceful hours rest by the side of the road after knocking over a couple of ogres. And random encounters raise the same issue.
This issue - of the problems that arise from using an in-fiction concept ("rests take 1 hour") to manage what is essentially a pacing question ("I need this many enemies to be fought before the first short rest, or else the resource attrition won't work like I wanted it to") - has been discussed at great length in relation to 5e. (And before then, going back to discussions around 5 minute workdays and nova-ing.) The only solution that is generally offered is that the GM should pour on time pressure such that the players won't want to rest because if they rest they'll lose. In effect that's the 13th Age approach, only cloaked in a veneer of in-game logic. (And how thin that veneer is, and how likely to be seen through if the same pattern keeps recurring, is a major topic of discussion on those threads.)
I don't really recommend those threads, as they rarely cut through to the analytical nub: which is that the fundamental question is "How much do I want to ration short rests relative to the events of the game"? And if the answer is "no short rest between every moment of fisticuffs" then you need to focus on that, and then think about what technique you are going to use to enforce the rationing. Getting caught up thinking that the means (the technique for rationing short rests) is an end in itself - which is a frequent feature of those threads - is just a distraction from the real discussion, which is the feasibility of different techniques for diffrent playstyles and different campaign set ups. (Eg as I've said, a time-extension to short rests might work for a dungeon campaign, but probably won't help with the two ogres on the road.)