An HPA is configured on CPU but never scales up, even under load. Which prerequisite is most likely missing?
- A The metrics-server is not running or the pods have no CPU requests defined (correct answer)
- B The cluster autoscaler is disabled
- C The deployment lacks a PodDisruptionBudget
- D The service is of type ClusterIP
Shortcut: HPA computes utilisation as a percentage of the CPU request. Without resource requests, or without metrics-server supplying metrics, utilisation cannot be calculated and scaling never triggers.