As a player, one of the main problems I have in an encounter is figuring out when to deploy daily and other resources. Sometimes it's easy, such as when a character really needs healing to stay in the fight (you spend the surge), or when targets are perfectly positioned for that area-affecting daily power. But I often horde my non-encounter resources waiting for either a dire circumstance or a perfect opportunity to use them.
Nod. The best guideline, IMX, is "if you see a good (not perfect) opportunity to make use of a daily and accomplish something worthwhile, use it." If you wait for desperation you'll make poor use of your dailies - using dailies with great riders when the rider doesn't matter because you're desperate to do that extra couple dice of damage, that kind of thing. If you make good use of your dailies, you're less likely to get to the point of desperation.
The phenomenon I often see is dailies vs surges. Early in the day, players will horde dailies, making fights take longer and taking more damage in the process, until they start running low on surges. Then they'll bust out dailies to try to end fights before their few remaining surges get used up.
Another thing to consider is that even at 1st level, as a (typical) /party/ you have 5 dailies. You can afford to let one character expend a daily in every fight - even if a fight doesn't seem /that/ important or overwhelming, if one daily can be used very well in it, and save you all a number a surges, it may well be worth it.
The same goes for action points (though their milestone-renewable) and consumables (which tend to have 'expiration dates' in the sense that, as you level, the consumable drops off precipitously in value.
Another issue in 4e, again, perhaps only for me, is figuring out if you can take a certain enemy. Some knowledge or insight checks might convince the DM to give you some intel, but it's rarely gonna forecast the outcome of the fight. And 4e fights are often properly engineered so that the party is at the brink of losing, sometimes several times, before finally winning a fight. This is intentional: PCs have fewer HP than monsters, but many more ways to recover HP. Plus, it makes the fight much more fun.
4e's approach to adventure design is /very/ different from prior eds. If a DM has really bought into it, this issue basically doesn't exist. Any challenge you in encounter will be statted out based on what it represents to you. If a creature is far out of you league, it won't have combat stats, you won't be able to fight, rather, you'll have a skill challenge to escape it or negotiate with it or it'll be a plot device to make something else happen. Ideally, guessing at what critters you can 'take' mechanically (or memorizing monster manuals and metagaming the same question) isn't an element of 4e play, at all. Rather, if a creature is overwhelming or contemptible, that will be represented in the story, rather than via pointless combat statistics.
But, as in 3e, the DC on monster knowledges really don't help. Monsters that you should avoid like the plague because they're way out of your league you won't even be able to identify, while you'll have encyclopedic knowledge of obscure monsters that should be beneath your notice. Prettymuch backwards from what would be really helpful for playability. Unbeatable uber-monsters should be legendary, everyone knows to avoid or placate them.