A production GraphQL service should expose operational health information appropriate for its infrastructure. Liveness indicates that the process is functioning, while readiness indicates that the instance is prepared to receive traffic.

Readiness may depend on required dependencies being available, although health checks should be designed carefully so a temporary dependency problem does not create cascading failures across the entire deployment.