I have a problem with ram in k8s.
My scenario 800CCU for 5 minutes
I testing performance Keycloak on K8S with configuaration:
replicaCount: 5
resources:
requests:
cpu: 1000m
memory: 2048Mi
limits:
cpu: 3000m
memory: 4096Miautoscaling:
enabled: true
minReplicas: 5
maxReplicas: 15
targetCPU: “75”
targetMemory: “75”
And, my problem are is the requests cannot be divided equally among each pod so some pod always run max 4GB Ram and it has to restart.
Is there a way to free up ram on the pod?