The current implementation of our GraphQL API was an experiment that for many reasons did not get much adoption. It runs on a separate port and we were never able to fully integrate into the main code. For that reason, it has added too many dependencies that are useless for the vast majority of our deployments. It is the source of annoying issues like #10714 causing unnecessary load and worries.

We believe that supporting GraphQL is still on our long term plans, but we will have to refactor the implementation with the new architecture we are designing for high scalability.

With that in mind, we are planning to remove the current implementation from the code base and keep the PR #8158 as a reference for future development as an auxiliary Apolo base project (probably).