It's not great, but just to make sure you're not mis-reading that.. after 1000 hours at 2.5V in a 60°C environment, it lost at most 30% of its capacitance. You should get much better performance at room temperature and below the rated 2.5V. Note that the tolerance is also pretty broad.. -20% to +80%, itself. If you need a more accurate and stable capacitance, these probably shouldn't be your first choice :) If you just want to use them as intended - as a reasonably reliable power source that charges up in no time and can then keep a system running (e.g. a microcontroller + some sensors in standby) for a good while or just as an emergency back-up in intermittent power loss situations - it shouldn't be a problem.

yes, 2.5V is the absolute maximum.

It might work to put a diode in series with the capacitor's charge path to drop the voltage a little bit. You could probably find a Schottky diode with the right forward voltage drop pretty easily. This does mean, of course, that whatever's being powered by the capacitor has to be able to run off of 2.5 V (and even less—capacitors do not maintain voltage at all while discharging).

If the strand is fine enough, sure.
The ESR is ~0.1 ohm, so to couple well to the wire, the wire will need to have higher resistance than that. And it stores ~4 joules energy, so to glow red-hot you'll need a pretty low-mass wire (0.39 J/gK for copper, so you have about 10 gK to work with to get your temperature rise of around 650 K). Happily for you, both of these restrictions point towards using as fine a copper strand as you can find. 40 gauge or finer, it looks like. I think vacuum-cleaner mains cables tend to be made of fairly fine stranded wire, though maybe not as fine as 40ga.
Have fun!