"Wet" electrolytics began as surge limiters to protect motors and transformers. Large wet units will reliably absorb huge energy short-term.
Today we only have "dry" e-caps and no short-term reserve.
Yes, the rise is semi-exponential so a little more voltage is a lot more current. A sane designer will always keep a significant safety factor from the maker's rated voltage.
> relatively small, limited (2mA) current, the capacitor stops with the voltage increase at nearly 30% above the rated Vdc. The energy dissipated seems too little to make the capacitor heat
That's a reasonable result. While a large cap "may" have over 2mA leakage at rated voltage, 99% of production will have much less.
Considering the -wide- range of e-caps, it is pointless to cite a current without the uFd or Voltage or can size.
Taking wild numbers: 2mA in a "450V" cap over-volted to 600V (+30%) means 1.2 Watts of heat in the can. For a 1"d 1"h (25x25mm) can of about 3 square inches exposed surface, 0.4W per sq.in. or about 40 deg C temperature rise. Hardly "cool".
The next problem is that leakage increases with temperature. Reasonably double for each 10 deg C rise. If the 2mA is solidly limited, then current will not increase, voltage will decrease, so power and temperature decreases. The system is stable. If current can increase, it will. That increases power and temperature and heat and leakage... run-away.
I'm not over-concerned about damage to thin foils. The oxide layer is nearly mono-molecular. While foils and etchings have become ultrathin, still any practical foil is far-far thicker than the oxide layer, lots of reserve. (Not the same as Metalized Film construction.) I would however worry about the byproducts of oxide destruction contaminating the electrolyte. (Tho I worry more about chloride contamination built into the cap at the factory.)