The Horizontal Pod Autoscaler (HPA) automatically scales the number of pods based on CPU utilization, memory usage, or custom metrics. Scaling Targets per Service Service Min Max CPU Target Memory Target gateway 2 6 60% 70% identity-api 2 4 70% 80% catalog-api 2 10 70% 80% commerce-api 2 8 70% 80% engagement-api 2 4 70% 80% web 2 6 60% 70% Scaling Behavior Configure stabilization windows to prevent flapping — scale up quickly (30s window) but scale down slowly (300s window) to handle traffic fluctuations.