How to understand the calculation method of the kubernetes.io/batch-cpu value on a machine

[KrWMdev]

What happened:
How to understand the calculation method of the kubernetes.io/batch-cpu value on a machine, and what are the sources of factors that affect the changes in the kubernetes.io/batch-cpu value?
I tried to read the source code, but I didn't understand (for example, adding some deployments without actual load increases the allocated CPU requests on the machine, but the kubernetes.io/batch-cpu does not increase).
batchAllocatable := nodeAllocatable * thresholdPercent - podUsage(HP priority classes higher than Batch ) - systemUsage
https://github.com/koordinator-sh/koordinator/blob/main/pkg/slo-controller/noderesource/framework/README.md

Environment:

• Koordinator version: - v1.4.1
• Kubernetes version (use kubectl version): v1.28.2
• docker version: cri-dockerd + docker 19.03.14
• OS (e.g: cat /etc/os-release): rhel7.9
• Kernel (e.g. uname -a): Linux 4.19.9-1.el7.elrepo.x86_64 ✨ Add NodeMetric API #1 SMP Thu Dec 13 10:52:03 EST 2018 x86_64 x86_64 x86_64 GNU/Linux

Anything else we need to know:

[saintube]

@KrWMdev Please refer to the Resource Model design doc: https://koordinator.sh/docs/architecture/resource-model