If you search hard enough, you'll find life curves for very many capacitors, including ones at volts varied away from rated.
Effects of increasing applied voltage also varies by type/construction method.
I used to work in tantalum cap' R&D and the insulator/dielectric, which is tantalum pentoxide, is formed electrolytically, in the ones that I worked on, at around twice rated volts, so they would go short-circuit at around double rated volts. Between rated and forming volts, they get progressively more leaky (higher leakage current - Li).
Failure mechanism in things like film cap's is going to be way different, but no doubt discussed online in many and various places.
I have no idea if they are now or ever were available generally, but some grades were aged using Weibull aging, which can produce cap's with truly massively extended rated life as the principle allows early failures to be eliminated from any one batch, so average expected/rated life of the remainder is increased. Too complicated to go into here, but again, no doubt, lots online if intrigued.
With any cap' with an immobile, non-volatile electrolyte, potential life is limitless. The common quote that "they dry out" applies to some but by no means all cap' designs. It does not apply to electrolytic tantalums as there is nothing in them to "dry out", they are 100% solid state, indeed, moisture ingress just makes them electrically leaky to some degree.