pipeop threshold, classif.log_reg, pipeop scale

[mmiche]

I am currently uncertain whether the following setup makes sense or not:
The mlr3book example on threshold tuning uses a decision tree as classifier, which operates the same way, irrespective of whether the features have been scaled or not (as far as I know). If logistic regression was the classifier instead of rpart, is scaling the features still recommended. My main goal in this case is to tune the threshold. My question is, if scaling makes sense, how would I have to link the pipeops threshold and scale correctly with the log_reg classifier?