A GraphQL userContext query returned data: null with errors, failing in multiple regions — the kind of generic error that invites a wrong fix at the resolver layer.
02
Task
Find why the query failed, rather than patching symptoms.
03
Action
I inspected the configuration files from inside the running pod, then found the pod logs were full of 'Not allowed by CORS' — pointing the investigation at CORS/origin configuration, and confirmed it reproduced across regions.
The resolver was fine; the request was being rejected upstream by CORS before it could resolve.The real error, in the pod logstext
# GraphQL response (misleading):
{ "data": null, "errors": [{ "message": "Internal server error" }] }
# Pod log (the truth):
WARN o.s.web.cors.DefaultCorsProcessor - Rejected request: Not allowed by CORS
=> fix the origin/CORS config, not the resolver
04
Result
Pinpointed a configuration root cause behind a generic GraphQL error, avoiding a wrong fix at the resolver layer.
Confirmed the per-region config difference and corrected the origin/CORS configuration.
Reinforced: read the pod logs for the real error before changing application code.