I’m not going to give a list of examples because list questions are off-topic here, and because I don’t have a list handy. Instead, I’m going to talk about the problems with Epic Spells.
Also, caveat, I have not really spent a lot of time with Epic Spellcasting; I saw the problems that they had, and wrote them (and all of the Epic Level Handbook) off as useless, a long time ago.
Open-Ended
The first, and obvious, problem with Epic Spells is that they are extremely open-ended. You combine different factors together, and theoretically you plug in the appropriate numbers and figure out how difficult a thing this is to do. The problem with this is that it is a system with an astounding number of possible combinations, and there is absolutely no way for a designer to consider even a small fraction of them specifically. That means all of the rules that he or she designs for the system are based on generalities, and there’s a mathematical guarantee that those generalities are not going to fit quite a few combinations.
So on some level, the system cannot be balanced. And I’ve seen very thorough, competent, and mathematically-inclined designers attempt to do so: this has tended to result in absurdly complicated rules that still have plenty of cases where they break.
DM Judgment
All of Dungeons & Dragons, and most likely all of any RPG, depends on the DM’s judgment to function. However, because of the above fact, Epic Spells tend to be purely a matter of DM judgment, since the system itself is all-but-useless. This results in more work for the DM, and potentially greater risks for the campaign if he makes a poor judgment.
Because the rules are so close to useless, a DM is close to better off ignoring them entirely and deciding on the fly whether or not a given thing works. The problems with that can be myriad (arguments about what works or not, yet more work for the DM, can be hard-if-not-impossible for the player to plan ahead and rely on something, etc.), but ultimately are probably fewer than the problems of the actual Epic Spellcasting system itself.
Mitigation
Ultimately, though, the real problem with Epic Spellcasting is mitigation. Epic Spellcasting can actually be broken “both ways,” that is, underpowered or overpowered, depending on one’s (ab)use of mitigation.
No Mitigation: Epic Spellcasting is useless
The DCs could probably be hit if you were really optimizing Spellcraft, but the times or other problems involved would make most Epic Spells completely useless. This includes almost every statted Epic Spell. The power of 9th-level spells is enormous, and without considerable mitigation, Epic Spells just do not keep up.
Heavy Mitigation: Epic Spellcasting can do anything
You can stack mitigation factors ad infinitum, and reduce the costs of your Epic Spell to next-to-nothing if you try hard enough. In particular, getting a lot of followers helping to cast the spell is not a difficult thing for a properly-motivated Epic-level spellcaster.
Meanwhile, you can combine the abilities of Epic Spells to break the (already few) limits of spellcasting in powerful ways. You can literally create an Epic Spell to do more-or-less exactly what you want in each case. Mitigate it a lot, and you can have it ready and cast very quickly.
Where do we draw the line? Epic Level Handbook has no help for us.
It’s basically impossible to say in a general way where the line between “completely useless” and “absurdly overpowered” is. Partially because 9th-level spells are already more-or-less “absurdly overpowered” as it is, and therefore Epic Spellcasting is supposed to be better than that somehow without breaking the game (impossible, I’d argue, because 9th-level spells already have). So of course, the Epic Level Handbook does not actually have any useful guidelines for us, and we return to my previous point about pure DM adjudication.
