❯ apps · Running

Dashboard du cluster avec Homepage + Glances Cluster Dashboard with Homepage + Glances

Construction d'un dashboard homelab unifié avec Homepage pour les liens de services et Glances pour les métriques système en temps réel — déployé sur Kubernetes avec découverte automatique des services. Building a unified homelab dashboard with Homepage for service links and Glances for real-time system metrics — deployed on Kubernetes with auto-discovered services.

· 5 min de lecture 5 min read ·
k3s-clusterdashboarduxmonitoring
🇫🇷 Lire cet article en français

🔄 Update — June 2026. Homepage was decommissioned on 2026-06-28 and replaced by a hand-built static hub. This article describes the architecture as it existed before that change — kept for historical reference.

A single entry point

With 15+ exposed services, navigating the homelab without a centralised dashboard becomes tedious. Homepage solves this with a clean web interface, configurable in YAML, that gathers every service together with its real-time status.

Glances complements Homepage by exposing detailed system metrics (CPU, RAM, network, disk) through a REST API that Homepage consumes.

Homepage

Homepage is configured entirely through ConfigMaps. The main configuration defines the services grouped by category:

# ConfigMap homepage-services
services.yaml: |
  - Infrastructure:
    - ArgoCD:
        href: https://argocd.olig-homelab-2025.win
        icon: argocd.png
        description: GitOps controller
    - Longhorn:
        href: http://10.0.0.11:30080
        icon: longhorn.png
        description: Distributed storage (NodePort)
  - Development:
    - Gitea:
        href: https://gitea.olig-homelab-2025.win
        icon: gitea.png
        description: Self-hosted Git
    - code-server:
        href: https://code.olig-homelab-2025.win
        icon: vscode.png
        description: VS Code in browser
    - Claude WebUI:
        href: https://claude.olig-homelab-2025.win
        icon: anthropic.png
        description: Claude Code via web UI

Homepage supports widgets that display live metrics pulled from the services themselves (ArgoCD status, Longhorn volumes, uptime…) through their APIs.

Glances

Glances exposes system metrics through a REST API on port 61208. Homepage can consume that API to show CPU/RAM/network directly in the dashboard.

Glances is deployed in server mode:

containers:
- name: glances
  image: nicolargo/glances:latest
  args: ["-w", "--disable-plugin", "docker"]
  ports:
  - containerPort: 61208

Lessons learned

RBAC for auto-discovery: Homepage can discover Kubernetes services automatically by reading the annotations on Ingresses. This requires a ServiceAccount with the appropriate permissions on the services, endpoints and ingresses resources.

Icons: Homepage ships with a built-in icon library for popular services. For custom services, host the icons in the homepage-icons ConfigMap or reference an external URL.

In hindsight: this dashboard did its job until it was replaced in June 2026 (see the update at the top of the article). Most of the value it provided — a single entry point to the homelab services — was carried over to its successor.

k3s-clusterdashboarduxmonitoring