We are kick starting the POC to make our application headless using DX API V2. We have few questions regrading its architecture and capability.

1) Does DX API v2 follows HATEOAS architecture. 

• Does UI call a ‘common service’ after every screen is painted? Or will the response from first screen give pointers to the second screen?

2) Do we have any documentation that details each and every elements in the response structure specially UI metadata. Pega API help doesn't give much insights on this. We need to know what controls and actions are supported along with metadata structure for those controls.

3) Do we have any best practices to follow before considering DX API V2 adoption.

4) Does DX APIv2 maintains user session and locking all along. If yes, how it achieves that.