Is the latter not the case? I'm genuinely asking here, as this is a point of view expressed by my tech lead (though in our case it's a graph DB vs an existing MongoDB setup).
Absolutely unrelated. You can think of GraphQL as a standardized glorified RPC that calls arbitrary functions that return some data. Whether the source of data is an RDBMS, a bunch of REST microservices, DynamoDB, redis, SQLite, a flat JSON file, Neo4j, or a D12 dice doesn't really matter.
The "graph" part is if your arbitrary data is actually somehow related, you can traverse those relationships in one request instead of having to do many calls in a waterfall.
Not especially, no. It likes trees, or things that can easily and naturally be represented as trees, best, as far as I've ever been able to tell. Granted, I suppose, that's a kind of graph.
The "graph" part is if your arbitrary data is actually somehow related, you can traverse those relationships in one request instead of having to do many calls in a waterfall.