OtterScale Dashboard

Kubernetes multi-cluster management UI — SvelteKit + Connect RPC.

Architecture

graph LR
    Browser -->|HTTP/SSR| SvelteKit[SvelteKit]
    Browser -->|Connect RPC<br/>same-origin + x-proxy-target| SvelteKit
    SvelteKit -->|OAuth 2.0 / OIDC| Keycloak[Keycloak]
    SvelteKit -->|Session / refresh lock| Redis[Redis]
    SvelteKit -->|Proxy + Bearer| API[OtterScale API]
    API --> K8s[Kubernetes Clusters]
    API --> Prometheus[Prometheus]

    subgraph Connect RPC used by Dashboard
        API --- LinkService[LinkService]
        API --- ResourceService[ResourceService]
        API --- RuntimeService[RuntimeService]
    end

Loading

🚀 Quick Start

# Clone
git clone https://github.com/otterscale/dashboard.git && cd dashboard

# Configure
cp .env.example .env   # edit .env

# Dev
pnpm install && pnpm dev

# Production (Docker)
docker build -t otterscale/dashboard .
docker run -p 3000:3000 --env-file .env otterscale/dashboard

⚙️ Configuration

ENV_VAR	Default	Description
PUBLIC_WEB_URL	http://localhost:3000	Public URL (required)
PUBLIC_HARBOR_URL	http://localhost:5000	Public Harbor URL (required)
API_URL	http://localhost:8299	OtterScale API endpoint (required)
REDIS_URL	—	Redis connection string (required)
KEYCLOAK_REALM_URL	—	Keycloak realm URL (required)
KEYCLOAK_CLIENT_ID	—	OAuth client ID (required)
KEYCLOAK_CLIENT_SECRET	—	OAuth client secret (required)

Features

• Multi-cluster K8s management — scopes, resources, RBAC
• Storage orchestration — Ceph pools, OSD, NFS, SMB, object gateway
• VM lifecycle — create, manage, console via xterm.js
• AI/ML model serving — LLM deployment & inference proxy
• Prometheus monitoring — built-in metrics & dashboards

License

AGPL-3.0