Which statement differentiates a Headless service from a standard ClusterIP service?

Which statement differentiates a Headless service from a standard ClusterIP service?

• A. A Headless service does not allocate a cluster IP, allowing direct endpoint resolution.
• B. A Headless service automatically creates an external load balancer.
• C. A Headless service uses a single virtual IP for all endpoints.
• D. A Headless service is only used for external access.

Answer: (A)

In Kubernetes, a Service can expose a set of pods either through a single virtual IP or directly through the endpoints themselves. A standard ClusterIP service allocates a cluster IP and uses that IP to route traffic to one of the backend pods, providing internal load balancing. A headless service, by contrast, does not allocate a cluster IP, meaning there is no single virtual IP to balance against. Instead, DNS for the service returns the actual pod endpoints, allowing clients to resolve and connect directly to each pod. That direct endpoint resolution is what the statement is describing. Since no cluster IP is created, clients can discover and reach the individual pods behind the service rather than going through a single load-balanced IP. This behavior is useful for certain stateful or database-like workloads where direct pod access or custom client-side logic is needed. Other options don’t fit because headless services do not automatically create external load balancers, they do not rely on a single virtual IP, and they aren’t restricted to external access—headless services are often used for internal service discovery and direct access to endpoints.

In Kubernetes, a Service can expose a set of pods either through a single virtual IP or directly through the endpoints themselves. A standard ClusterIP service allocates a cluster IP and uses that IP to route traffic to one of the backend pods, providing internal load balancing. A headless service, by contrast, does not allocate a cluster IP, meaning there is no single virtual IP to balance against. Instead, DNS for the service returns the actual pod endpoints, allowing clients to resolve and connect directly to each pod.

That direct endpoint resolution is what the statement is describing. Since no cluster IP is created, clients can discover and reach the individual pods behind the service rather than going through a single load-balanced IP. This behavior is useful for certain stateful or database-like workloads where direct pod access or custom client-side logic is needed.

Other options don’t fit because headless services do not automatically create external load balancers, they do not rely on a single virtual IP, and they aren’t restricted to external access—headless services are often used for internal service discovery and direct access to endpoints.