Data structure (JSON) good practice for services

I will explain firstly a usecase: Let us assume that I have a service e.g. a server which generates some data in form of a JSON object.

Server 1: {"id":1234, "timestamp":"2020-11-12"}

There is another server (server 2) which fetches the data occasionally. Maybe each second but maybe each 12 hours. Server 2 also adds new data to the JSON.

Server 2: {"id":1234, "timestamp":"2020-11-12", "my-new-data":"someHash"}

So now the question:

1. As I know how the resulting data will look like, shall I define it consistent from the beginning at Server 1?

--> Server 1: {"id":1234, "timestamp":"2020-11-12", "my-new-data":null}

Cons (while defining structure at each server the same):

• This way I may have huge JSON objects and while fetching very often I have more traffic, which could be in fact avoided.
• If JSON object changes, each server needs to be updated
• Dirty architecture?

Pros (while defining structure at each server the same):

• Consistency
• Cleaner architecture?

Which way would you choose and why? Do you differentiate over different usecases - e.g. small service, big service?