this happens only with one of the clusters (the one that I have updated the kubernetes version), the other one, managed with the same rancher instance, works just fine