feat: add Kubernetes integration with monitoring POD resources

[ntheanh201]

Issue Type

• Improvement/feature implementation

Runtime Environment

• Kubernetes cluster
• Operating system and version: [e.g. Ubuntu 20.04 LTS / Windows 10 Build 19043.1110]
• Terminal emulator and version: [e.g. GNOME Terminal 3.36.2 / Windows Terminal 1.8.1521.0]
• Python version: [e.g. 3.8.2 / 3.9.6]
• NVML version (driver version): [e.g. 460.84]
• nvitop version or commit: [e.g. 0.10.0 / 0.10.1.dev7+ga083321 / main@75ae3c]
• python-ml-py version: [e.g. 11.450.51]
• Locale: [e.g. C / C.UTF-8 / en_US.UTF-8]

Description

• This PR introduces Kubernetes integration to nvitop.
• Allow monitor Kubernetes resources by POD

Motivation and Context

• Closes: [Feature Request] 是否考虑出支持K8S的版本？ #169

Testing

Images / Videos

[XuehaiPan]

Hi @ntheanh201, thanks for the contribution. I wonder if you could add a separate package (use a new folder nvitop-kubernetes) for Kubernetes support. Then you can add a kubernetes extra here (merge the latest main):

nvitop/setup.py

Lines 75 to 87 in 5a4f1d5

	extra_requirements = {
	'exporter': ['nvitop-exporter'],
	'prometheus': ['nvitop-exporter'],
	'lint': [
	'ruff',
	'pylint[spelling]',
	'xdoctest',
	'mypy',
	'typing-extensions',
	'pre-commit',
	],
	'cuda10': ['nvidia-ml-py == 11.450.51'],
	}

---

For the CLI monitor, you can subclass the GpuProcess class in the new nvitop_kubernetes module, then reuse the nvitop.cli with the following patch:

from nvitop.tui import Device
from nvitop_kubernetes.tui import ModifiedGpuProcess

Device.GPU_PROCESS_CLASS = ModifiedGpuProcess

For POD information in the CLI, you can prepend the information to the process command-line instead of adding a new column. Adding a new column could be too complex because there will be too many places to change (e.g., process sorting support, horizontal scrolling support (← → arrow key)).