We currently worked with an API that had to scale to serve multiple regions and millions of users. Introducing caching for such numbers is a common and battle-tested way to reduce costs. With services like Stellate, this is no longer impossible even when using GraphQL.

That is great news, however, also poses some challenges. We need to pay attention to how structure our types so we can fully utilize caching.

How to rethink some GraphQL schema designs so we can have the best of the data-driven paradigm of GQL and have amazing caching capabilities at the same time?