We are running a stateful database application running on Kubernetes cluster with a minimum 3 nodes required. I have configured Virtual pod autoscaling for the pods using auto mode. Now the issue is When I scale up the cluster with 3 pods, One of the pods is unable to schedule due to Insufficient memory while sufficient memory is available in the cluster node. I want all 3 pods running all time.
My concern is why the pod is not scheduled while memory is enough on the node.
Cluster information:
Kubernetes version: 1.19.14-gke.1900
Screenshot attached here.
