Kubernetes service discovery is a fundamental aspect of container orchestration that facilitates communication between microservices. This system enables applications to locate and interact with each other dynamically, promoting flexibility and scalability in distributed architectures.

At its core, Kubernetes employs DNS-based discovery. Each service is assigned a DNS entry in the format ..svc.cluster.local. This approach allows pods to resolve service locations without hardcoding IP addresses, enhancing portability and ease of configuration.

In addition to DNS, Kubernetes automatically injects environment variables into pods. These variables contain the IP addresses and ports of active services, providing an alternative method for service discovery. While simple, this method can be less flexible in dynamic environments.

For more advanced use cases, Kubernetes exposes an API that applications can query directly. This method offers greater control and real-time updates but requires more complex implementation.

Label selectors play a crucial role in associating services with their respective pods. This mechanism allows for dynamic service composition, adapting to changes in the cluster's state.

Kubernetes also incorporates readiness probes to ensure that only pods capable of handling requests are included in the service's endpoint list. This feature enhances the reliability of service-to-service communication.

For scenarios requiring direct pod-to-pod communication, Kubernetes offers headless services. These services provide DNS entries for individual pods rather than a single cluster IP, enabling more granular routing.

Lastly, Kubernetes provides mechanisms for discovering external services, either through ExternalName type services or manual endpoint configuration.

This comprehensive approach to service discovery in Kubernetes underpins its ability to manage complex, microservices-based applications effectively. By abstracting the complexities of network communication, Kubernetes allows developers to focus on building scalable and resilient systems.