Your approach is good but what you should mention is that in DDD there is a concept of Aggregates and a necessary rule for it to have only one transaction on one Aggregate (see here Page 7: https://www.dddcommunity.org/wp-content/uploads/files/pdf_articles/Vernon_2011_2.pdf). The reason why this is so important is that people build architectures calling tons of lambdas to keep their objects in sync while missing the point of event driven applications and create non scalable systems

Sounds like a really good library. I like the non annotation approach

Sandro Volpicella I think before using feature flags and shrink everything into smaller PRs a team should always try to focus on the business increment for the user and make that small as possible in the first place. Sure, this is a hard thing but from my perspective a „big“ feature comes always with „big“ testing efforts regardless of the PR sizes and makes it anyway hard to review