Geena, a cloud architect, uses a master component in the Kubernetes cluster architecture that scans newly generated pods and allocates a node to them. This component can also assign nodes based on fac

Kube-controller-manager runs controller loops that regulate cluster state (e.g., replication controllers, endpoint controllers), but it does not directly assign pods to nodes.

Kube-scheduler continuously monitors the API server for unscheduled pods and selects an optimal node for each based on criteria including available CPU/memory resources, data locality, affinity/anti-affinity rules, and hardware or policy restrictions. Once a node is selected, the scheduler updates the pod definition via the API server so the kubelet on that node can start the pod.

Kube-apiserver is the front-end REST interface for the Kubernetes control plane that processes and validates API requests; it does not make scheduling decisions.

Etcd is a distributed key-value store used to persist all cluster configuration and state data; it has no role in evaluating or assigning pods to nodes.