senior54 zgłoszeń

Jak działa Horizontal Pod Autoscaler?

HPA automatycznie skaluje liczbę replik na podstawie metryk. Domyślnie CPU/Memory (Metrics Server). Custom metrics przez Prometheus adapter. Algorytm: desiredReplicas = ceil(currentReplicas * (currentMetric / targetMetric)). Min/max replicas. Cooldown periods.

Więcej pytań z Kubernetes