Chapter 9 addresses (to some extent) logging and monitoring, and goverance - which is a lot to do with how the model should be managed in life....

I've worked in projects where the model was ungoverned and went wrong and no one noticed for a long time... and that caused damage. I also got called in to sort out a project where the team retrained the model every week... and every week they overfitted it on new data. I think knowing what the models should do, being able to say that they are doing that and then having a clear way of deciding what to do if they aren't (ie. someone in charge) is the base of maintaining them... what's your pov though?