QLDB, of course, isn’t free. But if you’re an AWS customer, it makes sense to use QLDB instead of spending more money and time building, running, and maintaining a custom data auditing system that will have its own bugs and vulnerabilities.
Aside from maturity and cost concerns, resistance to adopting QLDB seems to stem from the fact that teams must at least partially solidify their object model. Teams that choose NoSQL may particularly resist this requirement. Schema stability is a reflection of the maturity of a project. Data models for new projects are constantly in flux. Schema stability is required for both production readiness and building useful reports. While data lakes store snapshots that can be used for auditing purposes, only products like QLDB can store and analyze immutable historical data without custom coding.