The core issue is that 4E monsters deal relatively low damage per attack.
If resistance 5 would only protect against 25% or less of the incoming damage, the potions wouldn't feel as over-powered (even if the cost-benefit analysis stays the same, i.e. if the ratio between your healing surge value and the resisted damage stays the same)
As it is now, a potion can easily protect against 50% or more of each incoming typed attack, and often 100% against level-appropriate ongoing typed damage.
One solution is to mix it up: adding monsters that aren't dealing damage of the same type, adding monsters that also make untyped attacks. But this feels more like a cop-out to me (not to mention the biggest cop-out of all: using multiple-type damage attacks!)
What I am taking away from the discussion so far is mostly that the resistance values are simply too high given the game's damage ratings. Other solutions feel a bit simple ("ban!") or a bit complicated ("add a new roll!").
I could go with resist 2/4/6 or "they only work once each round" but I'll probably try to deal with the most egregious usage: stopping ongoing damage.
If those potions work on regular attacks but can't stop ongoing damage they will still be useful, but not so completely overwhelmingly good as they are now. And it makes for a simple houserule easy to remember.
And they won't trivialize dragon breath.