For me the clue is in the delayed rev-counter sweep issue. On the video, it showed an immediate sweep on switch-on, as it should. But then a second one a few seconds later! That should not happen. The only thing that could cause that, to my mind, is either an intermittent break in the ignition circuit somewhere, or a voltage on the cluster motherboard that is on the margins.
The problem is, as i see it, there's no easy way to test the actual voltage on the motherboard unless you have a complete test setup with cluster and ecu connected off the bike. You cannot get to the connector pins to test voltages when everything is connected. I had a similar problem years ago on my first T1050, and there is a cluster pins diagrammatic somewhere in the stickies.
In the end I found in my case (by hours of testing and the wiring diagram in hand) that the fixed +12V feed was not there - and the reason was that a fuse in the fusebox that looked OK actually was not. Only testing with a meter showed that it was blown, even though it looked OK.
But you have a worse problem, an intermittent connection somewhere, the hardest of all faults to trace. Seems to me that it has to be down to the power feeds (either the fixed live or the switched live, or the earth connection) causing the problem. Repeated sweeps suggest an intermittent power feed break - whereas the earlier long delay suggests that the voltage is building up slowly enough to start a sweep.
It all points to a connection somewhere that is corroded, could be internal and invisible even, and supplying power intermittently. I know you've tested those microswitches, but have you thought of simply resoldering them on the board (in which case you could replace them anyway).
And of course what caused the corrosion. Has the bike gone from a damp to a very dry environment? Or vice versa? Humidity in garages has been known to cause problems in the UK.