Erratic Residuals

I am simulating a hedge (2.2x25x0.35m: https://www.simscale.com/workbench/?pid=5058303832479289449&rru=f824a20d-6a0b-433c-b527-1dffdca4faf3&ci=7b20f0c6-5a65-4fea-8f9f-f0a65a70bd69&ct=CONVERGENCE_PLOT&mt=SIMULATION_RESULT ) or de box (5x5x5m, SimScale Login ), And I see erratic behavoir of the residuals (mainly in epsilon and k):

For hedge:

For the box:

The graph of the Ux though looks nicely converging:
E.g. in the hedge project (Probe points plots→ MP → Ux)

So how important is this erratic behavoir in Residulas (in my case I am not that interested in k, my interest is Ux)?

Thanks for your help.