I’m not sure it could be a hardware issue unless there is a soldering problem or poor quality resistor/cap in the circuit. The input 3.3V line could have an issue? The circuit is dead simple. I reckon it has to be a firmware problem with the sampling speed, resolution or noise of the PT1000 sensor on the micropin.
I have the same hardware as you and noticed my rapido2 with pt1000 had a similar behavior where it wasn’t stable and varied +/-2 or 3 degrees. However, i had my configured for 4700ohm w/o the jumper.
I found a similar post in the duet forum: PID Tuning RTD PT1000 from the 1LC | Duet3D Forum