I think it's just two sides of the same coin. There's definitely an art to making it consumable to end users. It's even an engineering challenge to open these data lakes to end users without killing the database (think of a dev trying to show load average on a graph for all running servers over last 3 years and there's 5000 servers). You can't really blame them and realistically they might have a project where they need that data. And then there's data crazy people who make 400 graphs of various metrics and only 20 get actively used. It can lead you down red herring paths and all kinds of chaos as an SRE, and also blow the budget.

Anyway, I understand what they are saying and unfortunately I have no book recommendations to solve these problems. In larger companies it can be much more of a challenge than actually ingesting data, especially in the observability space.