docs: add k8s guide and Llama-3.2-3b-Instruct example-run

[sallyom]

I've been running in K8 with the job added in this PR, maybe it's useful to document as an example for everyone. Also, adding a simple analyze_benchmarks.py script to analyze the results of a GuideLLM run. Here's an example showing the plots generated by the analyze_benchmarks.py script run with output from a minikube/Llama3.2-3B-Instruct run.

[sjmonson]

Apologies for the late review. We were waiting for our official container image to land; see comments below about replacing your included dockerfile with the official image.

[sjmonson]

We now have our own image. Utilize that and the built-in support for environment variables (see /deploy for more details).

Suggested change

	- name: guidellm
	# TODO: replace this image
	image: quay.io/sallyom/guidellm:latest
	imagePullPolicy: IfNotPresent
	securityContext:
	allowPrivilegeEscalation: false
	capabilities:
	drop:
	- ALL
	runAsNonRoot: true
	seccompProfile:
	type: RuntimeDefault
	args:
	- benchmark
	- --target=$(TARGET)
	- --data=$(DATA)
	- --rate-type=sweep
	- --model=$(MODEL)
	- --output-path=/app/data/llama32-3b.yaml
	env:
	# HF_TOKEN is not necessary if you share/use the model PVC. Guidellm needs to access the tokenizer file.
	# You can provide a path to the tokenizer file by passing `--tokenizer=/path/to/model`. If you do not
	# pass the tokenizer path, Guidellm will get the tokenizer file(s) from Huggingface.
	- name: HF_TOKEN
	valueFrom:
	secretKeyRef:
	key: HF_TOKEN
	name: huggingface-secret
	- name: TARGET
	value: "http://llm-d-inference-gateway.llm-d.svc.cluster.local:80/v1"
	- name: DATA_TYPE
	value: "emulated"
	- name: DATA
	value: "prompt_tokens=512,output_tokens=128"
	- name: MODEL
	value: "meta-llama/Llama-3.2-3B-Instruct"
	volumeMounts:
	- name: output
	mountPath: /app/data
	- name: guidellm
	image: ghcr.io/neuralmagic/guidellm:latest
	imagePullPolicy: IfNotPresent
	securityContext:
	allowPrivilegeEscalation: false
	capabilities:
	drop:
	- ALL
	runAsNonRoot: true
	seccompProfile:
	type: RuntimeDefault
	env:
	# HF_TOKEN is not necessary if you share/use the model PVC. Guidellm needs to access the tokenizer file.
	# You can provide a path to the tokenizer file by passing `--tokenizer=/path/to/model`. If you do not
	# pass the tokenizer path, Guidellm will get the tokenizer file(s) from Huggingface.
	- name: HF_TOKEN
	valueFrom:
	secretKeyRef:
	key: HF_TOKEN
	name: huggingface-secret
	- name: GUIDELLM_TARGET
	value: "http://llm-d-inference-gateway.llm-d.svc.cluster.local:80"
	- name: GUIDELLM_RATE_TYPE
	value: "sweep"
	- name: GUIDELLM_DATA
	value: "prompt_tokens=512,output_tokens=128"
	- name: GUIDELLM_MODEL
	value: "meta-llama/Llama-3.2-3B-Instruct"
	volumeMounts:
	- name: output
	mountPath: /app/data

[sjmonson]

Also note --data-type is deprecated, I removed it in the suggestion. Additionally --rate-type was not configurable so I added it as an env.

[sjmonson]

Drop this file since we now have an official container.

[sjmonson]

Suggested change

> **📝 NOTE:** [Dockerfile](./Dockerfile) was used to build the image for the guidellm-job pod.

[sjmonson]

Suggested change

	> This eliminates the need for Guidellm to download the files from Huggingface.
	> This eliminates the need for GuideLLM to download the files from Hugging Face.

[sjmonson]

There are a bunch case issues with the GuideLLM name. Please fix to be constant with the rest of our documentation.

Suggested change

	## Run Guidellm with Kubernetes Job
	## Run GuideLLM with Kubernetes Job

[sjmonson]

Also please rebase on main for latest CI. Development / build has an issue and will always fail, but all others should pass. You can run ruff check --fix docs/guides and ruff format docs/guides to auto fix the majority of linting issues.