That looks to me to be bad data. Is the VE.Direct cable from the shunt to the GX running past any active gear that could be spewing a lot of RF? things like microwaves, poor quality phone chargers, ballast from fluorescent lights, that kind of thing.
Other things to try;
- swap the VE.Direct cable. The signal rate over ve.direct is very low, and its a near bulletproof serial signal protocol that can run over almost anything, but there still could be a fault with the cable or someone jammed it in a door hinge or stuck a staple through it.
- Change the ve.direct port on the GX
- Victron shunts generally never need calibration, but maybe someone tried to calibrate it and completely messed it up??? (calibration is covered in the shunt manual )
- sometimes relying on a chassis negative that looks fine but isn’t actually attached to the chassis can yeild strange results.
- Do you have aluminium bus bars? weird results
- Read @pwfarnell post here