I'm on my way to think about optimize our Terragrunt repositories. We're using it for years, following best practices. Our organization is looking like the official infrastructure and infrastructure-modules registries.
We wrote our own modules using plain Terraform with always in mind to stay DRY. But for some kind of module we sometimes almost rewrote the official modules, even taking inspirations from some of them. Also It's sometime hard to be up-to-date, without refactor our modules again and again. It's a pure loss of time.
We're trying to consume the official registry modules by just using the tfr:// protocol But by doing this we're loosing some of the flexibility what can be offer when we wrote our own modules (keep a quick examples : our AWS tagging plan) or just because modules can be multiple ressources at the same time (example : include IAM roles with an EC2 instance module).
So I'm looking for some guidance about this topic. Maybe some feedbacks regarding this. What's the best way to optimize both our code and our time.