What it is
The kubelet turns a Pod's CPU and memory requests and limits into cgroup v2 files: cpu.weight from the CPU request, cpu.max from the CPU limit, memory.max from the memory limit, a QoS class that decides where the Pod sits in the cgroup tree, and an oom_score_adj that decides who the kernel kills first under memory pressure. This tool applies those rules to the YAML you paste and lists every file with the rule it came from, for the Kubernetes version, cgroup driver and runtime you select.
When to use it
Use it to check what a limit really does before you set it: how 500m becomes 50000 100000, why 1 CPU is weight 39 on the pod cgroup but 100 in the container cgroup with a recent runc, why a Burstable Pod is killed before a Guaranteed one, what Memory QoS changes, and which cgroup paths to cat on a node to confirm.
Gotchas & limits
This is a model of what the kubelet writes, not a view of a real node: it follows the kubelet source on the master branch as read on 2026-10-20 and the rule table says which sources were compared. Pod-level resources are not modelled. The container directory name depends on the runtime. Check the exported commands on a real node before relying on a value.