TL:DR noisy pin signal, due to interference between MCU timer and PWM.
I know, different FW, but maybe issue is the same since it’s the same board. What I am planning to do: move Servo Pin to another group, maybe 1, raise mechanically the probe after homing , and try a bit longer print.
Update: changed servo pin to group 1 (1.28 actually) . Now I am 1h45m onto printing same g-code (this time without filament) and so far the probe pin is stowed. I’ll update tomorrow , when the print finishes.
After 2.5 hours was ok. Then Klipper restarted with nulnulnulnulnul entry in the log, so I suppose in klipperian language means that my SD card is dying? Anyway, consider the probe issue as solved
Rpi PSU looks reliable-ish and worked fine just a day before with the same g-code. The SD card I found in my drawer , who knows where did I took it from