I have two applications running inside the same Kubernetes cluster. One is a backend Service exposing a REST API, and the other is a client application (running as a separate Deployment) that should call this API.
When I try to access the backend service using the full Cluster DNS name (for example, http://my-backend-service.default.svc.cluster.local:8080) from inside the client pod, the request fails with a “stumble game host not found” or connection error. However, if I use the pod IP directly, it works.
I have already checked that the backend deployment is running, the service is created with the correct selector, and DNS is enabled in the cluster, I can also ping other built-in-service like Kubernetes.default.
What could prevent my application pod from resolving or reaching the service name via DNS? Are there specific network policies, RBAC settings, or service configuration issues that might block this service discovery?