The CI is failing everywhere, because the following is failing everywhere

from sklearn.datasets import fetch_openml
fetch_openml(name='miceprotein', version=4)

even tho the dataset exist: https://www.openml.org/d/40966