Steps to reproduce:

1. I have got a kubernetes cluster of 3 master nodes and 3 worker nodes.
2. Install kubernetes zabbix template "Kubernetes nodes by HTTP"
3. Open most recent data for memory/cpu limit of specific node. Example: kube.node.limits.memory[k8s-worker-p02]

Result:
Should see the exact same limit values from this command:

kubectl describe node k8s-worker-p02 | grep -A 10 "Allocated resources"

Expected:
But there is a slight difference in the value.

See the images below.

MEMORY: 34640 MB is different than 35.61 GB
CPU: 16 is different than 25