what is the expected average latency for server which exposes graphql api ?

[vusenko]

Hello everyone,

I have considered graphql-mesh as gateway server for existing project which represents itself a distributed system, i.e. set of microservies which works and exposes rest api(which implements openapi spec) and graphql. The graphql-mesh looks like a silver bullet for us, but during the simple check I've observed the latency in avg for 300 ms for a server which exposes graphql api. Is it something which expected as overhead ? and what is the average latency should be expected in this case?

For test we took https://api.spacex.land/graphql/ and simply compared the response time for direct calls and over the graphql-mesh.

example of config:

sources:
  - name: SpaceX
     handler:
      graphql:
        endpoint: https://api.spacex.land/graphql

query for testing

{
  launchesPast(limit: 10) {
    mission_name
    launch_date_local
    launch_site {
      site_name_long
    }
    links {
      article_link
      video_link
    }
    rocket {
      rocket_name
      first_stage {
        cores {
          flight
          core {
            reuse_count
            status
          }
        }
      }
      second_stage {
        payloads {
          payload_type
          payload_mass_kg
          payload_mass_lbs
        }
      }
    }
    ships {
      name
      home_port
      image
    }
  }
}

Where is the original response time

and response time over the graphql-mesh