What is the best design for read optimized replica?

I need to design a scalable, read optimized access to a set of data we replicate from outside.

For some of the data queried we have to keep in mind that it is not just data but code and data. If you take the example of hierarchies you will see that the plain data e.g. a relational DB will not be very helpful. A service might need all children of a node and ends up implementing recursions. if we simply replicate the data into every service you we will end up with a lot of code duplication and many answers to the same question.

On top of my mind, im planning to use CQRS. Can you guys please help to validate for the above requirement CQRS is good approach or any other design is available to create scalable read replicas?