Boeing pilots have stated on this website how the oxygen bottle can be switched off yet residual pressure in the delivery lines can produce a convincing test of mask function. I tried a similar thing in an Airbus and the same happens. Conclusion: test the mask delivery system thoroughly, ie a good, long press-to-test to make sure there is a continuous supply of gas behind the lines.
Studies have shown that even mild levels of hypoxia will produce a degradation of a pilot’s performance some time after the event, even after descending out of hypoxia territory. (Pilots were made hypoxic and then put in the sim to carry out simulated diversions following a rapid decompression. More errors were made in the hypoxic group than a normal, control group. I can’t remember the reference to the paper - sorry). I wouldn’t even hang around at 25k, but maybe that’s just me.
Gauge failure or system failure? I don’t know and I don’t care above 25K. How do you feel if the fuel gauges drop to zero but the engines are still running? I may suspect the gauge but I also feel nervous. I cope with nerves better on the ground than I do at altitude, but maybe that’s just me.
The “rules” of O2 at FL 310 and FL250 are not different. It’s the hazards that are different. The hazards of crashing a car at 30 mph are different to crashing at 80 mph. The smart money, however, always remains on not being in a position to crash in the first place. It sounds so easy to slip on a mask, twiddle a knob here, pull another knob there and it’s home in time for tea and medals. Maybe it is that simple, on your day. Unfortunately there is well documented evidence of trained pilots attempting such “simple” actions yet still ending up unconscious. This suggests it is not always going to be your day.
I’m not preaching or being judgmental, just stating what I’d do and why, which is get down and get it checked over. Others obviously wouldn’t. Good luck to them - and their passengers.