Glad you got the o-rings done successfully Paul - and its useful to know that the ball and retainer can be ejected if there is enough pressure in the system, so care needs to be taken to avoid this. I've only ever done the job with the pump removed from the car and the plug facing upwards during its removal so haven't run into this issue. As you say, replacing these two o-rings is a fairly easy job and in most cases these two o-rings are the only sources of leaks on the pump other than the inlet hose.
In theory having the bleed screw undone should have been enough to prevent the pressure you saw, but I'm usually a stickler for always de-pressurizing the whole system including suspension before doing hydraulic work. Certainly when I did the tank pipe mod and had the top out of the tank I had the suspension well and truly de-pressurized to avoid any leak back of oil through the now disconnected pipes, and indeed there was none, barely more than a few drips in the hour or so it was apart.
Fingers crossed the ride stays crash free!

Mine has been very consistently crash free, however it has made more obvious a couple of other minor issues - one is that the front ride height sometimes seems to "creep" higher during driving, enough that you can tell from your perspective out the drivers window that the car is "leaning back" slightly. When it does this the ride at the front becomes a bit "jiggly". Not harsh though. When I measure the ride height while stationary the front ride height is spot on, so it's something that is happening dynamically, possibly the manual override linkage is out of adjustment, I've never checked it.
BTW, I did notice a significant reduction (but not elimination) of ride crashiness when I had my pump out to replace those two o-rings - and this was well before I did the 10 minute mod. Whether the o-rings were admitting air on the intake strokes or whether there was a poor seal on the main inlet hose that I inadvertently fixed when refitting the pump I can't be sure, but there was a definite improvement in both ride quality and consistency of ride quality. Any sources of air ingress into the hydraulic pump, whether at the pump or before it are cumulative. Only when all of the air leaks and sources of air bubbles through the inlet are eliminated does the ride become truly crash free and consistently good.
Interesting to see that the ride on yours was still a bit crashy when the LHM level was 1.5 litres too low but was good once topped up - this tallies with my observation that some bubbles do still seem to get through to the pump (despite the modification) when the oil level in the tank is low, (due to the suspension being fully raised) but not when the oil level is normal (normal ride height) or maximum. (minimum ride height) It does seem that there needs to be enough depth of oil in the tank to form a buffering layer to allow any bubbles/turbulence to disperse - if the level is too low there is no buffer layer and the bubbles/turbulence are down close to the pump inlet depth allowing the bubbles to be sucked in.
Regarding the temperature warning light coming on when first starting and going away when you rev the engine a bit - check your coolant level !!

The temperature warning light above the temperature gauge is also used for low coolant level, measured in the expansion bottle depending on the engine.
When my V6 was using a little bit of coolant I'd notice this sometimes too - the warning light would flicker on and off a few times when starting then go away after 30 seconds or so - the heat and circulation of the coolant in the engine was just enough to satisfy the level sensor after it had been running for about a minute. The coolant was only an inch or two too low in the expansion bottle, that's all it took for the warning light to flicker on.