edge-cli is the unified Go command-line entry point for the Edge-Computing-LLM
GitHub organization.
It coordinates two current layers:
k3s-nvidia-edge: infrastructure for a local Linux + k3s + NVIDIA GPU host.llm-observability-stack: LLMOps workloads, Helm charts, Open WebUI, Ollama, OpenTelemetry, Prometheus, Grafana, and related services.
The organization also publishes
gguf-observability,
a read-only evidence companion for deployed GGUF model runtimes. It consumes the
status produced by these layers; edge-cli does not install it as another
cluster layer.
Dashboard presentation is owned by the Helm-provisioned Grafana dashboards in
llm-observability-stack; there is no separate browser repository or
browser-side Kubernetes credential path.
The CLI supports automatic NVIDIA/CPU selection. It does not currently install AMD, Intel, or Apple Silicon accelerator runtimes.
The organization has separate repositories for infrastructure and LLMOps
workloads. edge-cli provides one installable control plane so operators can run
preflight checks, installs, validation, status, logs, and safe uninstall flows
without treating each repository as a disconnected CLI application.
Future repositories can be added as new layers in internal/platform and new Go
modules under internal/modules.
Build from source:
go mod tidy
go build -o edge ./cmd/edge
sudo install -m 0755 edge /usr/local/bin/edge
edge versionDefault config path:
~/.edge-cli/config.yaml
Create it:
edge config init
edge config showDefault configuration:
repos:
k3sNvidiaEdge: /media/waqasm86/External1/Waqas-Projects/Project-Linux-Kubernetes-Nvidia/Project-Edge-Computing-LLM/k3s-nvidia-edge
llmObservabilityStack: /media/waqasm86/External1/Waqas-Projects/Project-Linux-Kubernetes-Nvidia/Project-Edge-Computing-LLM/llm-observability-stack
cluster:
kubeconfig: ""
defaultNamespace: llm-observability
gpu:
vendor: nvidiaRepository paths can also be overridden per command:
edge install infra --repo-path /path/to/k3s-nvidia-edge --yes
edge install observability --repo-path /path/to/llm-observability-stack --yesedge doctor
edge status
edge install infra --yes
edge validate infra
edge install observability --yes
edge validate observability
edge install all --yes
edge logs
edge repo list
edge repo doctoredge install all --accelerator auto --yes
edge statusauto deploys the NVIDIA infrastructure layer when nvidia-smi detects a
working GPU. Without NVIDIA hardware it skips the toolkit and GPU Operator,
installs or validates basic k3s, and selects values.cpu-k3s.yaml.
On the validated GeForce 940M profile, the application smoke test targets
qwen-1-8b-chat-q4-k-m-local; CPU and general profiles retain their configured
Gemma model defaults.
Explicit modes are useful in automation:
edge install all --accelerator nvidia --yes
edge install all --accelerator cpu --yesedge install all runs:
- Host accelerator detection.
- Linux/k3s installation and accelerator-specific infrastructure setup.
- Basic k3s validation, plus NVIDIA and CUDA validation in NVIDIA mode.
llm-observability-stackHelm install.- Observability validation.
- Next command and URL hints.
An empty local k3s cluster with only the default k3s system components is a
valid starting point. edge install infra is responsible for preparing the
NVIDIA runtime and GPU Operator layer before any GPU-backed LLM workload is
installed.
edge install infra uses the local k3s-nvidia-edge project as the
infrastructure source. It validates the repo path, checks Linux and NVIDIA
preconditions, installs required host packages and NVIDIA Container Toolkit when
not skipped, installs k3s when missing, installs the GPU Operator through the
bundled Helm chart by default, and validates GPU readiness with a CUDA pod.
Useful commands:
edge install infra --yes
edge install infra --skip-k3s-install --yes
edge install infra --skip-toolkit-install --yes
edge validate infra
edge uninstall infra --yesedge install observability uses the local llm-observability-stack chart. It
validates the infra layer first, builds Helm dependencies, installs the chart,
waits for Ollama, Open WebUI, and OpenTelemetry Collector, then runs validation.
For GPU profiles, edge-cli does not allow bypassing infra validation and
forcibly keeps gpu-operator, nvidia-device-plugin, and dcgm-exporter
disabled in the observability chart. Those components belong to
k3s-nvidia-edge.
On single-GPU laptops, run edge validate infra before installing
observability. After Ollama is running and reserving nvidia.com/gpu: 1,
observability validation checks base readiness without launching another CUDA
pod that would contend for the only GPU.
Useful commands:
edge install observability --profile geforce-940m-k3s --yes
edge install observability --values custom.yaml --set ollama.enabled=true --yes
edge validate observability
edge logs --tail 200
edge uninstall observability --yes --keep-namespaceUninstall commands require --yes. Observability uninstall does not delete user
data outside Kubernetes resources. Use --keep-namespace to retain the
namespace. Infra uninstall keeps k3s unless --k3s is explicitly set.
edge uninstall observability --yes --keep-namespace
edge uninstall infra --yes
edge uninstall all --yesRun the broad checks first:
edge doctor
edge repo doctor
edge statusCommon issues:
- Missing
nvidia-smi: install or repair the NVIDIA driver on the host. - Missing
runtimeclass/nvidia: validate the GPU Operator and NVIDIA Container Toolkit setup withedge validate infra. - Helm release missing: run
edge install observability --yes. - Pods not ready: run
edge logsand inspect the namespace withkubectl get pods -n llm-observability -o wide.
go test ./...
go build -o edge ./cmd/edgeThe project is 100% Go. It does not create Bash scripts. External platform
commands such as kubectl, helm, apt-get, systemctl, k3s, and
nvidia-smi are executed through Go's os/exec.