MiaAI-Lab · Trending

MiaAI-Lab/sparkDash

sparkDash ⚡ — Multi-DGX Spark Monitoring Dashboard

View on GitHub Website / Demo Back to listing

678stars 138forks 26contributors JavaScript

README.md

# sparkDash ⚡ — Multi-unit monitoring dashboard for NVIDIA DGX Spark

macOS (darwin) SSH collectors authored by <a href="https://github.com/MikeGibbsOnyx">C. Michael Gibbs</a> and <a href="https://x.com/nyxvoss_ai">Nyx Voss</a> (Onyx AI Labs) — <a href="https://github.com/MikeGibbsOnyx/sparkDash/commits?author=MikeGibbsOnyx">commits</a>. sparkDash by <a href="https://github.com/MiaAI-Lab">Mia's AI Lab</a>.

sparkDash is a real-time web dashboard for one or more **NVIDIA DGX Spark (GB10)** machines in a single browser window. It streams GPU, CPU, unified memory, storage, network, and local LLM metrics — and lets you add, edit, reorder, or remove Sparks from the UI without restarts or code changes.

It also supports **non-Spark units**: any Linux machine with an NVIDIA GPU (e.g. a workstation with a dedicated RTX/L-series card) can be added as a **dedicated GPU host** and monitored the same way via SSH and `nvidia-smi`. For these units the dashboard correctly separates **RAM** (system memory) from **VRAM** (discrete GPU memory).

Monitored units run **Linux**, **macOS**, or **Windows** (a Windows PC with an NVIDIA GPU — or an AMD-iGPU Windows host like a Strix Halo, whose GPU usage comes from Windows' GPU engine counters; see [Windows PCs](#windows-pcs)).

### LLM Prompt Showcase

<a href="./.github/llm-showcase.mp4">Watch the MP4</a>

---

## Table of contents

- [Latest version changelog](#latest-version-changelog)
- [Features](#features)
- [LLM server presets & backend auto-detection](#llm-server-presets--backend-auto-detection)
- [ComfyUI monitoring](#comfyui-monitoring)
- [Hermes Agent monitoring](#hermes-agent-monitoring)
- [Tailnet monitoring](#tailnet-monitoring)
- [Tailscale Serve (HTTPS)](#tailscale-serve-https)
- [Mobile PWA (install on a phone)](#mobile-pwa-install-on-a-phone)
- [Glance integration](#glance-integration)
- [Quality bench](#quality-bench)
- [Prometheus](#prometheus)
- [RoCE / RDMA monitoring](#roce--rdma-monitoring)
- [Windows PCs](#windows-pcs)
- [Full changelog](./CHANGELOG.md)
- [Quick start](#quick-start)
- [Architecture](#architecture)
- [Tech stack](#tech-stack)
- [Repository layout](#repository-layout)
- [REST API](#rest-api)
- [Configuration](#configuration)
- [Security](#security)
- [Scripts](#scripts)
- [How it works](#how-it-works)
- [Contributing](#contributing)
- [License](#license)
- [Acknowledgements](#acknowledgements)

---

## Latest version changelog

### Version 2.40.0
- **Hide any Overview panel**: Settings has switches for Recent activity and for each of the four tiles (Fleet decode, Fleet power, Memory in use, Tokens today), next to Fleet energy and LLM token totals.
- **LM Studio and Ollama** are detected by their own endpoints, so they are no longer labelled SGLang or vLLM; LLM panels match their port, and the Add wizard and LLM panels have LM Studio / Ollama / TensorFold port presets.
- **Windows PCs with an AMD iGPU (Strix Halo)** show GPU use and memory from Windows' GPU counters. Windows setup docs now cover the `administrators_authorized_keys` gotcha.

### Version 2.32.0
- **LLM panel**: an engine that only SSH can reach (a remote unit behind a NAT that forwards just SSH) is now monitored through an SSH forward, with a **via SSH** chip; TensorFold prefill tok/s counts only computed tokens, so prefix-cache hits no longer read as 100k+.
- **Decode benchmark**: live generation tok/s is measured from the run's own streams, so it no longer shows 0.
- **Health**: a finding is held for 90 s, so a reading sitting on a threshold does not flap and spam Activity.
- **iPhone app**: the header clears the status bar and the home-screen icon is the logo. Dropping a Spark after dragging it in the sidebar no longer opens it.

### Version 2.31.0
- **RoCE / RDMA panel** for Sparks: link state and rate, traffic, drops, PFC and QoS settings and the RDMA counters per port, with two health findings (a link went down, packets being dropped). See [RoCE / RDMA monitoring](#roce--rdma-monitoring).
- **Time zone setting** for the Tokens, Energy and Activity charts, so the 24 h charts line up with your local time.
- **`/metrics`** now shows when each collector last returned real data, so a stuck GPU poll can be alerted on while SSH liveness still passes.
- Spark page: Unified memory, CPU and Models share one row; CPU and RAM sit side by side on GPU hosts; the GPU history chart fills its panel; Windows PCs got a first real-hardware pass.

### Version 2.30.0
- **Windows PCs as units**: a Windows PC with an NVIDIA GPU can be monitored over its built-in OpenSSH server (GPU, RAM, CPU load, disks, network, shutdown). See [Windows PCs](#windows-pcs).
- **Mobile app**: sparkDash installs as a PWA over HTTPS (for example Tailscale Serve).
- **Monthly energy history** with a month/year view, **Prometheus `/metrics`** (opt-in), per-GPU graphs on multi-GPU hosts, a model picker in the decode benchmark, per-Spark `llmHost`, and **Restart sparkDash** from the UI after a fleet change.
- The Settings poll interval now really sets the polling rate; the Fleet energy card explains partial coverage.
- Fixes: "Frontend not built" after a Docker upgrade, the SSH tunnel for benchmarks, model launchers inheriting `PORT=5555`, and a token-protected remote bind showing a blank page. Everything since 2.0.0 is in the [CHANGELOG](./CHANGELOG.md).

### Version 2.0.0 — a new sparkDash
- **New shell**: sidebar, command palette (Ctrl/⌘ K), four themes (White, Light, Dark, OLED), smooth page transitions and real links everywhere (right-click, new tab).
- **Rebuilt Overview**: per-Spark cards with the model row, head/worker relations, a Model/System memory bar, and a launcher to load a model from the card.
- **Benchmarks section**: Decode, Prefill, Quality and the new **Tool Eval Bench** (trials, side-by-side compare, one-click upgrade).
- **Showcase** is now an in-app page with much smoother streaming; **Token totals**, **Fleet energy** and **Activity** get their own pages.
- Start and stop your own LLMs from the dashboard; live prefill tok/s during long prefills; Quality bench with GSM8K and MMLU.
- **Health findings** per Spark (hot GPU, low unified memory under 3 GB, Xid/OOM errors, slow links) and Activity events for them; **Clear / Reset** for Activity, Fleet energy and Token totals.
- **Mobile**: docked tab bar with iOS-style motion, a Stats tab (Token totals, Fleet energy, Activity) and a rebuilt Spark page.

Full history: [CHANGELOG.md](./CHANGELOG.md)

---

## Features

| Area | What you get |
|------|----------------|
| **Multi-unit** | Any number of units; each has a tabbed detail page plus a shared Overview |
| **Non-Spark GPU hosts** | Linux boxes with a dedicated NVIDIA GPU are first-class units: same `nvidia-smi` collectors over SSH, detected hardware summary, and separate **RAM** / **VRAM** panels. Detail page: GPU (left) + **RAM → Network → Storage** (right column); Overview cards show RAM and VRAM bars |
| **Windows hosts** | Windows machines with an AMD iGPU (Strix Halo and friends) are first-class units too: agentless PowerShell over OpenSSH, CPU/RAM/disks/network collectors, and GPU usage straight from the engine performance counters (reads like Task Manager). The Add/Edit wizard has a **Strix Halo / Windows host** unit type |
| **Live streaming** | WebSocket metrics with configurable poll intervals; central history store for sparklines across tab switches |
| **Local + remote** | Host metrics via sysfs/proc/`nvidia-smi`; remotes over SSH (key or password) |
| **LLM probe** | Auto-detects llama.cpp, vLLM, sglang, LM Studio, Ollama, ds4-server, EXL3, TensorFold, FreeToken, or q27; live decode/prefill tok/s from the engine's own counters where it publishes them (not LM Studio or Ollama); cached vs uncached prefill on ds4, llama.cpp, SGLang, and q27; **daily peak** history on the LLM card |
| **LLM server presets** | One-click **LM Studio :1234** / **Ollama :11434** / **TensorFold :8888** presets in the Add wizard's LLM-ports step, the "+ Add LLM port" panel, and each LLM panel's settings (they fill the port number — they install nothing) |
| **ComfyUI** | Opt-in probe: queue/jobs, progress, cancel, Open link, inventory, overview chip |
| **Hermes Agent** | Opt-in per unit: background update check (10 min), status badges, one-click or batch `hermes update` |
| **Tailnet** | Opt-in probe: flags a unit that is healthy on the LAN but off its tailnet |
| **Decode benchmark** | Multi-concurrency streaming decode tok/s; type picker (Structured / Prose / Code / JSON). Code is a different Python task per stream. Lab protocol (temp 0, thinking off); persisted last run. Remote units: LAN HTTP, or SSH tunnel to loopback. **Remote** button for an on-demand HTTPS/host:port target |
| **Prefill benchmark** | Context-size sweep (1k–300k) of prefill tok/s and TTFT; unique prefix per size; persisted last run. Same remote targeting as decode |
| **Quality benchmark** | Fixed, seeded quality suite (QA, reasoning, arithmetic, state tracking, GSM8K, MMLU, plus optional instruction following and long-context recall) at temperature 0; per-category and overall score; last 30 runs kept; paired compare with McNemar p-value. Same remote targeting as decode |
| **Prompt Showcase** | Full-page multi-terminal LLM streaming demo (up to 32 prompts) with live tok/s and copy-out |
| **LLM inference health** | KV cache free/pool (vLLM `cache_config_info` token and byte size; usage % when the pool size is absent), run/wait queue, TTFT/E2E/ITL p95, preemptions, prefix cache, MTP accept from Prometheus `/metrics` (vLLM and q27; q27 FIFO-queues so the Requests tile reads “N run” without a wait gauge) |
| **Multiple LLM ports** | Monitor several LLM servers on different ports simultaneously — each gets its own panel with independent backend detection and metrics |
| **GPU processes** | See the top GPU processes by VRAM usage directly in the GPU panel, including process name and memory allocation |
| **Multi-GPU hosts** | A dedicated GPU host with several NVIDIA cards reports each one: the header names every card, the GPU panel adds a block per card (throttle chip, usage and temperature sparklines, power, VRAM bar) and the API exposes `gpu.gpus[]`. The headline `gpu` numbers stay an aggregate of all cards, so Overview cards and alerts need no change |
| **Spark uptime** | System uptime displayed inline on each Spark header for at-a-glance availability |
| **Power controls** | Graceful shutdown (SSH host script). Wake-on-LAN is for dedicated GPU hosts whose NIC supports it. DGX Spark does not wake from a magic packet |
| **Spark roles** | **Head** / **Worker** / **Standalone** — worker label + head link; standalone can disable LLM monitoring; optional hide workers from Overview and tabs |
| **Unified memory** | GB10 128 GB LPDDR5X pool (~273 GB/s), GPU/CPU split, bandwidth via `nvidia-smi dmon`. Non-Spark hosts show discrete **VRAM** (nvidia-smi) and system **RAM** separately |
| **Health findings** | Small explainable rules over metrics already collected: GPU running hot (85 / 90 °C), GPU stuck in a low-power state, unified memory low (under 3 GB warns, under 1.5 GB critical), NVIDIA **Xid** errors and kernel **OOM kills**, several models loaded while memory is busy, a link slower than 1 Gb/s, and RoCE link / loss. Shown as chips on the cards and in full on the unit page; changes are logged to **Activity** |
| **Fleet pages** | **Token totals** (per model, hourly / daily history), **Fleet energy** (rolling 24 h / 31 d plus a permanent monthly archive, optional cost), and **Activity** (event log with filters). Each page can be cleared or reset from its **Clear / Reset…** menu |
| **RoCE / RDMA** | Per-port link state and rate, traffic, drops, PFC and QoS (trust mode, DSCP), and RDMA counters, read from sysfs / ethtool / `mlnx_qos` on the Spark (see [RoCE / RDMA monitoring](#roce--rdma-monitoring)) |
| **Windows PCs** | A Windows PC with an NVIDIA GPU is a first-class unit over the built-in OpenSSH server; an AMD-iGPU Windows host (Strix Halo) works too — its GPU usage comes from Windows' GPU engine counters (see [Windows PCs](#windows-pcs)) |
| **Mobile + PWA** | A phone layout with a docked tab bar (Overview, Sparks, Stats, Settings); installable to the home screen over HTTPS (see [Mobile PWA](#mobile-pwa-install-on-a-phone)) |
| **Time zone** | **Settings → Time zone** sets hour and day boundaries and labels on the Tokens, Energy and Activity charts, whatever device you view them on |
| **Prometheus** | Opt-in `GET /metrics` export (see [Prometheus](#prometheus)) |
| **Restart from the UI** | When the fleet changes and Energy needs a restart, a **Restart sparkDash** button restarts the Docker container (only offered when it can) |
| **Themes** | Dark, light, cool white, OLED — neutral palettes, persisted in `localStorage` |
| **Secrets** | SSH passwords AES-256-GCM encrypted; never in `sparks.json` or API responses |
| **Docker-first** | Single privileged container for host metrics; prod and dev Compose files |
| **Hot config** | Add / edit / remove / reorder Sparks from the UI with no process restart |

---

## ComfyUI monitoring

sparkDash can **optionally** monitor a [ComfyUI](https://github.com/comfyanonymous/ComfyUI) instance on each Spark — the same way it probes local LLMs, but focused on **jobs and queue**, not a second copy of GPU/RAM bars (those stay on the GPU / CPU panels).

### What is supported

| Capability | Details |
|------------|---------|
| **Opt-in per Spark** | `comfyMonitoring` (default **off**) + `comfyPort` (default **8188**) |
| **Any role** | Head, worker, and standalone can enable ComfyUI independently of LLM cluster role |
| **Liveness** | `GET /system_stats` — online, ComfyUI / PyTorch version, device type (cpu/cuda) |
| **Queue / jobs** | `GET /queue` — running + pending items; workflow **title**, **model/LoRA** filenames from the graph, footprint (**resolution · steps · sampler · batch · node count**) |
| **Progress** | Progress bar on the active job — Comfy WebSocket when events are available; otherwise elapsed / average-duration **estimate** |
| **Last finished job** | Status + duration via `/api/jobs` (fallback `/history`) |
| **Queue ETA** | Estimate from recent job durations × pending (+ progress remainder when known) |
| **Cancel / remove** | From the Comfy card: interrupt a running job or dequeue a pending one (`POST /api/sparks/:id/comfy/cancel`) |
| **Open ComfyUI** | One-click link to `http://{lanIp}:{comfyPort}` (LAN IP preferred so remote browsers do not hit localhost) |
| **Model inventory** | Checkpoints + LoRAs from `/models/*` (UI section only when at least one file is listed) |
| **Overview chip** | When monitoring is on: `Comfy · idle` / `run` / `Nq` / muted if unreachable |
| **Layout** | Under **Services**: primary LLM + Comfy side-by-side when both are enabled; collapsible **Resources** / **Services** sections |

**Not claimed:** true per-job VRAM (Comfy does not expose that cleanly over HTTP). Host GPU/VRAM remains on the GPU panel. Live step progress depends on Comfy broadcasting WS events; stock Comfy often scopes detailed progress to the client that submitted the prompt.

### How to enable (per Spark)

1. Open the Spark tab → **Edit** (pencil).
2. Enable **ComfyUI monitoring**.
3. Set **port** if needed (default **8188**).
4. **Save**.

The Spark page **Services** section shows the ComfyUI card. On Overview, a small Comfy chip appears for that unit.

**Connectivity Test** (in Edit) includes ComfyUI when monitoring is enabled.

### ComfyUI side requirements

- ComfyUI must be reachable from the **sparkDash server** on the probe host:
  - **Local Spark** (`isLocal`): sparkDash probes `127.0.0.1:{port}` (use Docker `network_mode: host` if the dashboard runs in a container).
  - **Remote Spark**: probe uses the Spark **LAN IP** (same as LLM probes).
- For **Open** from another machine’s browser, Comfy should listen on a reachable interface (e.g. `--listen 0.0.0.0`), not only loopback, and the Spark’s **LAN IP** must be set correctly in Edit.

### Config fields (persisted on the Spark)

| Field | Default | Description |
|-------|---------|-------------|
| `comfyMonitoring` | `false` | Probe ComfyUI and show the card / overview chip |
| `comfyPort` | `8188` | ComfyUI HTTP port |

### Related API

| Method | Path | Purpose |
|--------|------|---------|
| POST | `/api/sparks/:id/comfy/cancel` | Cancel a job (`{ "promptId": "<uuid>" }`) — interrupt running and/or remove from queue |

Env (optional): `COMFY_PORT` (default `8188`), `COMFY_PROBE_TIMEOUT_MS`, `POLL_INTERVAL_COMFY`.

---

## Hermes Agent monitoring

sparkDash can **optionally** monitor [Hermes Agent](https://github.com/nousresearch/hermes-agent) (nousresearch/hermes-agent) on each unit and run one-click updates for you over SSH.

### What is supported

| Capability | Details |
|------------|---------|
| **Opt-in per Spark** | `hermesMonitoring` (default **off**) in **Edit Spark** |
| **Auto update check** | Background `hermes update --check` over SSH (default every 10 min) — returns update availability + pending commits |
| **Status badges** | In the Spark header: `Hermes` (installed version), `Hermes not found` if the binary is missing |
| **One-click update** | **Update Hermes** button opens a dialog with live status, real pending commits, and release notes; **Update now** runs `hermes update` via SSH (non-interactive go) |
| **Update state** | Running / success / error surfaced live (button turns into a “Hermes updating… / failed” state) |
| **Batch update** | **Update Hermes** on Overview runs `hermes update` on every monitored unit, with a live per-unit progress bar |

### How to enable (per Spark)

1. Open the Spark tab → **Edit** (pencil).
2. Enable **Hermes Agent**.
3. **Save** — background checks start immediately.

The **Update Hermes** button appears in the Spark header/mobile action row; it turns warning-yellow with a commit-count badge only when an update is actually available. It also appears on Overview (batch) when at least one unit has Hermes enabled.

**Connectivity check note:** local units run the check as the **host user** (via `setpriv`/`nsenter`, never as container root); remote units run it over SSH. Either way, the logged-in user needs permission to read the Hermes repo.

### Side requirements

- **Hermes Agent must be installed on the target machine** — sparkDash only checks & updates; it does not install it. The binary is looked up in `~/.local/bin` and `/usr/local/bin`.
- SSH user must be able to run `hermes update --check` / `hermes update` non-interactively (key auth recommended).
- An update can take a few minutes (repo pull + dependency reinstall); a stale `*.lock` file from a crashed run is cleared before each attempt.

### Config fields (persisted on the Spark)

| Field | Default | Description |
|-------|---------|-------------|
| `hermesMonitoring` | `false` | Check/update Hermes Agent on this machine |

### Related API

| Method | Path | Purpose |
|--------|------|---------|
| POST | `/api/sparks/hermes/update-all` | Batch `hermes update` on every monitored Spark (Overview button) |
| POST | `/api/sparks/:id/hermes/check` | Force `hermes update --check` now |
| POST | `/api/sparks/:id/hermes/update` | Run `hermes update` in the background (202) |
| GET | `/api/sparks/:id/hermes/updates` | Update preview: latest release + installed version + real pending commits + resolved view |

Env (optional): `POLL_INTERVAL_HERMES` (default `600000` ms), `HERMES_UPDATE_TIMEOUT_MS` (default `600000` ms).

---

## Tailnet monitoring

Opt-in per unit (default **off**). Runs `tailscale status --json` on the host and shows a **Tailnet** card under Resources.

This closes a blind spot every LAN-based check shares, including sparkDash's own SSH liveness. When `tailscaled` loses its session with the coordination server, SSH/GPU/LLM can all stay healthy while the box is unreachable from off-LAN.

### What is supported

| Capability | Details |
|------------|---------|
| **Opt-in per unit** | `tailscaleMonitoring` (default **off**) in **Edit Spark** |
| **Off-tailnet detection** | `Self.Online` — the node's *own* view of the coordination server |
| **Reason, not just state** | Tailscale `Health` messages, backend state, tailnet IP, DERP relay, version, expired-key warning |

Asked of **each node about itself**. Peer state is never the verdict. The probe is read-only (`tailscale up` / `down` / `login` are never run).

### How to enable

1. Open **Edit Spark**.
2. Tick **Tailnet monitoring**.
3. Save. The Tailnet card appears under Resources.

### Host requirements

- `tailscale` CLI on the monitored host, and `tailscaled` running.
- Remote units: existing SSH. Local Docker: `nsenter` into the host mount namespace (same as `nvidia-smi`; `/host/proc` is already bind-mounted).

### Config fields

| Field | Default | Description |
|-------|---------|-------------|
| `tailscaleMonitoring` | `false` | Run `tailscale status --json` and show the Tailnet card |

Env (optional): `POLL_INTERVAL_TAILSCALE` (default `30000`), `TAILSCALE_PROBE_TIMEOUT_MS` (default `8000`).

---

## Tailscale Serve (HTTPS)

Tailscale Serve publishes the loopback-only dashboard over HTTPS to your tailnet — no ports opened, no reverse proxy, and a real certificate (`https://<node>.<tailnet>.ts.net`). It is also the easiest way to satisfy the HTTPS requirement for the [Mobile PWA](#mobile-pwa-install-on-a-phone), which browsers refuse to install over plain HTTP.

### One-time setup

1. **Enable HTTPS certificates and Serve for the tailnet** (admin does this once):
   open the approval link Tailscale prints on first use, or in the admin console enable
   **HTTPS** (MagicDNS → HTTPS Certificates) and **Tailscale Serve**.
2. On the sparkDash host, publish the dashboard (the dashboard must be running — Docker or `npm start`):

```bash
   sudo tailscale serve --bg --https=443 http://127.0.0.1:5555
   ```

If the container binds a specific address instead of loopback, proxy that address
   (check `docker logs sparkDash` for the `bind=` line), e.g.
   `sudo tailscale serve --bg --https=443 http://100.x.y.z:5555`.
3. Verify and get your URL:

```bash
   tailscale serve status
   # https://your-node.your-tailnet.ts.net  ->  proxy http://127.0.0.1:5555
   ```

Notes:

- The Tailscale hostname works with the default loopback bind — no `SPARKDASH_ALLOWED_HOSTS` entry needed (that list is only for custom reverse-proxy domains).
- Access is **tailnet-only**. Any device signed in to the tailnet (phone, laptop) can open the URL; nobody else can. For public exposure use `tailscale funnel` and an authenticated front door (see [Remote access](./docs/REMOTE-ACCESS.md)) — but prefer keeping it tailnet-only.
- The proxy forwards `/api/*` and `/ws` including WebSocket upgrades; nothing else to configure.
- To remove it later: `sudo tailscale serve --https=443 off`.
- Snap installs run Tailscale as root; if the plain command reports *Access denied*, use `sudo tailscale serve …`.

---

## Mobile PWA (install on a phone)

sparkDash is a **PWA**: from a supported browser it installs to the home screen and runs fullscreen like a native app, with its own icon and no browser chrome.

### What ships

| File | Purpose |
|------|---------|
| `public/manifest.webmanifest` | App name, `standalone` display, theme/background colors (`#0a0c0f`), icons |
| `public/icons/` | 192/512 px icons, `maskable` variants (Android adaptive icons), `apple-touch-icon.png` |
| `public/sw.js` | Service worker: caches the app shell; **never** caches `/api/*` or `/ws` (live data always comes from the network) |
| `src/components/shell/InstallPrompt.tsx` | In-app install banner (see below) |

### Requirements

- **HTTPS** is mandatory for install (or `localhost` for local testing). Use [Tailscale Serve](#tailscale-serve-https) or an authenticated TLS reverse proxy — a plain `http://<lan-ip>:5555` page cannot be installed.
- The manifest and icons are static files served by the same Express server (`dist/` after `npm run build`); no extra configuration.

### Installing

The app shows its own banner when opened in a browser it can be installed from (hidden once installed, or after dismissing):

- **Android / desktop Chrome:** an **Install** banner appears near the bottom of the page; tapping it opens the native confirm dialog. Also available in the ⋮ menu → *Install app* / *Add to Home screen*.
- **iPhone / iPad (Safari):** iOS allows no programmatic prompt; the app shows a one-time hint, then: **Share ⬆︎ → Add to Home Screen → Add**.
- The banner only appears on a fresh page load — reload once if you don't see it.

After installing, the icon sits on the home screen and the dashboard opens fullscreen with live data whenever the device can reach the server (on the tailnet, with Tailscale connected).

### Updating the installed app

Service worker caching: hashed `assets/*` are cache-first (immutable), `index.html` and `/` are network-first with a cache fallback. After deploying a new build, the installed app picks it up on the next open; no manual cache clearing is needed. If you ever need to force it, bump the `CACHE` version in `public/sw.js`.

---

## Glance integration

[Glance](https://github.com/glanceapp/glance) can show the cluster on a self-hosted dashboard with the community widget [`sparkdash-dgx-cluster`](https://github.com/glanceapp/community-widgets/tree/main/widgets/sparkdash-dgx-cluster) (contributed by [@linxichen](https://github.com/linxichen)). One `custom-api` card renders GPU temperature and usage, VRAM, generation tok/s, KV-cache usage, and uptime for a head node plus one worker, reading `/api/sparks/:id/metrics` directly:

```yaml
- type: custom-api
  title: sparkDash · DGX Cluster
  cache: 30s
  url: ${SPARKDASH_HEAD_URL}
  subrequests:
    node2:
      url: ${SPARKDASH_NODE2_URL}
  options:
    dashboardUrl: ${SPARKDASH_DASHBOARD_URL}
  template: |
    # paste the template from the widget's README
```

| Variable | Value |
|----------|-------|
| `SPARKDASH_HEAD_URL` | Metrics URL of one unit — `https://sparkdash.example.com/api/sparks/<id>/metrics` |
| `SPARKDASH_NODE2_URL` | Metrics URL of a second unit. For a single Spark, drop the `subrequests` map and the widget's second node block |
| `SPARKDASH_DASHBOARD_URL` | Dashboard base URL, used for the card's title link |

`/api/sparks/:id/metrics` is unauthenticated on loopback, the same as the rest of the dashboard. Read [Security](#security) before exposing it beyond a trusted network.

---

## Quality bench

The **Quality** button on the LLM card scores whatever model the port is serving on a fixed suite, so you can compare models, quantizations and KV-cache formats over time. Every item comes from a seeded PRNG and has a stable id: two runs are the same questions, paired item by item.

Every request is an OpenAI chat completion at temperature 0 with a fixed `seed`. Thinking is switched with the same flags the other benches use.

| Category | Items | Thinking | max_tokens | Scored by |
|----------|-------|----------|------------|-----------|
| **QA** | 150 (30 fixed + 120 generated: multiplication, add/subtract, string reversal, letter counts, sorting, binary, date offsets, LCM) | off | ≥ 64 | Accepted answer as a case-insensitive substring; counts, binary and LCM must be the last integer in the reply |
| **Reasoning** | 40 multi-step word problems | on | 8192 | Last `Answer: <jar>, <per box>` line in the final answer (not the reasoning) |
| **Arithmetic chain** | 40 ten-step chains (multiply, add, subtract, remainder, floor divide) | on | 12288 | Last `Answer: <number>` (markdown and thousands commas allowed) |
| **State tracking** | 40 token-transfer stories, 5 people, 25 events | on | 12288 | Last `Answer: <number>` |
| **GSM8K** | 200 grade-school maths problems, a fixed seeded sample of the public GSM8K test set (MIT) | on | 8192 | Last `Answer: <number>` (`$`, markdown and thousands commas allowed) |
| **MMLU** | 285 multiple-choice questions, 5 per subject across all 57 MMLU subjects, a fixed seeded sample of the public test set (MIT) | off | 512 | Last `Answer: <letter>` (or a bare letter) |
| **Instruction following** | 40 prompts with 2–4 verifiable formatting rules (case, length, bullets, paragraphs, keywords, endings…) | off | 1024 | Each rule is checked by code; an item passes only when every rule holds. Off by default so older overall scores stay comparable |
| **Long-context recall** | Items per size (default 2) at 8k–256k (default 32k, off by default) | off | 512 | 16 animal codes spread through filler text; 4 are corrected near the end. Passes when all 16 latest codes come back; overwritten codes are counted as *stale* |

The overall score is the mean of the category percentages. Long sizes that do not fit the model context (from the live probe or `/v1/models`) are skipped and listed.

The dialog shows live progress, the per-category table (mean completion tokens and how many replies hit `max_tokens` for the thinking categories, keys found and stale answers for long recall), and a collapsible per-item table with reply excerpts. Pick an earlier run under **Compare with** to see both scores, how many replies were identical, how many items only one run got right, and the exact two-sided McNemar p-value. When p ≥ 0.05 the row reads *difference within noise*. **Copy results** copies a plain-text summary.

The last 30 runs per unit are kept in `config/quality-bench-history.json`, with the label, model id, settings, and per-item pass/fail and reply hash. One quality run per unit at a time; it cannot overlap a decode or prefill bench or a showcase. A full default run is about 755 requests, and the thinking categories can take a while on slow models.

Quality bench by Mia's AI Lab.

---

## Prometheus

Opt-in (default **off**). **Settings → Prometheus metrics** serves every unit's metrics at `GET /metrics` in the Prometheus text format (0.0.4), so Prometheus, VictoriaMetrics or Grafana Agent can scrape sparkDash and keep history beyond what the dashboard shows. While the setting is off the path answers `404`.

`/metrics` goes through the same auth as the REST API: open on a loopback install, and on a remote bind with `SPARKDASH_TOKEN` set it needs the token as a bearer header.

```yaml
scrape_configs:
  - job_name: sparkdash
    scrape_interval: 15s
    static_configs:
      - targets: ["sparkdash.lan:5555"]
    # Only when SPARKDASH_TOKEN is set on a non-loopback bind:
    authorization:
      type: Bearer
      credentials_file: /etc/prometheus/sparkdash.token
```

The exporter reads the snapshot sparkDash already holds, so scraping adds no SSH or nvidia-smi calls; values are as fresh as the poll interval. Names follow node_exporter / DCGM-exporter practice: base units (bytes, seconds, °C, watts, 0–1 ratios), `_total` only on counters. Every series carries `unit` (the unit id), `unit_name`, `kind` (`spark` / `host`) and `role` (`head` / `worker` / `standalone`). Memory and disk sizes are collected in MiB and converted to bytes.

| Family | Type | Extra labels | Notes |
|--------|------|--------------|-------|
| `sparkdash_up` | gauge | | 1 when the unit answered its last liveness check. An unreachable unit exports `sparkdash_up 0` and nothing else |
| `sparkdash_uptime_seconds` | gauge | | Host uptime |
| `sparkdash_collector_last_success_timestamp_seconds` | gauge | `collector` | Unix time of the last poll of each collector (`gpu`, `cpu`, `ram`, `memory`, `network`, `storage`, `llm`, `roce`) that returned real data, not zeroed defaults. A stuck collector stops advancing while `sparkdash_up` stays 1, so alert on `time() - sparkdash_collector_last_success_timestamp_seconds > 60`. Absent until a collector has succeeded |
| `sparkdash_gpu_info` | gauge | `gpu`, `name`, `uuid` | Always 1. `gpu` is the nvidia-smi index for an NVIDIA card; an AMD card (amd-smi) is its index plus 1000, so `gpu="1000"` is the first AMD card. Per-card chart history uses the same number (`gpu.1000.*`) |
| `sparkdash_gpu_utilization_ratio` | gauge | `gpu` | 0–1 |
| `sparkdash_gpu_temperature_celsius` | gauge | `gpu` | |
| `sparkdash_gpu_power_watts`, `sparkdash_gpu_power_limit_watts` | gauge | `gpu` | |
| `sparkdash_gpu_memory_used_bytes`, `sparkdash_gpu_memory_total_bytes` | gauge | `gpu` | GB10: the GPU's share of / the whole unified pool |
| `sparkdash_gpu_throttled` | gauge | `gpu`, `reason` | 1 while clocks are limited for `thermal`, `power` or `hw` |
| `sparkdash_memory_available_bytes` | gauge | | Headroom for GPU work: unified-pool `MemAvailable` on a GB10, free VRAM on a discrete GPU |
| `sparkdash_cpu_utilization_ratio` | gauge | | 0–1 |
| `sparkdash_cpu_temperature_celsius` | gauge | `sensor` | A GB10 has no CPU package sensor; `sensor="acpitz"` says it is a board zone |
| `sparkdash_ram_used_bytes`, `sparkdash_ram_total_bytes` | gauge | | |
| `sparkdash_network_receive_bytes_per_second`, `…_transmit_bytes_per_second` | gauge | `interface` | Interfaces hidden in sparkDash are skipped |
| `sparkdash_disk_used_bytes`, `…_available_bytes`, `…_total_bytes` | gauge | `mount`, `device` | Devices hidden in sparkDash are skipped |
| `sparkdash_llm_up` | gauge | `port`, `backend`, `model` | One per LLM endpoint (head and standalone units) |
| `sparkdash_llm_generation_tokens_per_second`, `sparkdash_llm_prefill_tokens_per_second` | gauge | `port`, `backend`, `model` | |
| `sparkdash_llm_kv_cache_usage_ratio` | gauge | `port`, `backend`, `model` | Where the backend reports it |
| `sparkdash_llm_requests_running`, `sparkdash_llm_requests_waiting` | gauge | `port`, `backend`, `model` | Where the backend reports it |
| `sparkdash_llm_generated_tokens_total`, `sparkdash_llm_prompt_tokens_total`, `sparkdash_llm_cached_prompt_tokens_total` | counter | `port`, `backend`, `model` | The engine's own lifetime counters — they reset when the engine restarts, which `rate()` handles |

Unknown is left out rather than written as zero: a failed GPU read, a sensor that does not exist, or a field the LLM backend does not publish produces no sample.

---

## RoCE / RDMA monitoring

Spark units with RDMA devices (the CX7 ports) get a **RoCE / RDMA** panel on their page, with no switch integration and no root. Per port: link state and rate, MTU, RX/TX traffic, interface errors and drops, PFC priorities, QoS trust mode (PCP or DSCP) and the DSCP map, global pause settings, and the RDMA counters (out of buffer, sequence errors, ACK timeouts, ECN-marked packets, CNPs sent and handled) plus ethtool discards, CRC errors and pause frames, with the change since the last sample.

Everything comes from the Spark itself: `/sys/class/infiniband`, `/sys/class/net`, `ethtool -S` / `-a` and `mlnx_qos -i`. The fast part (link, counters, traffic) is read about every 5 s; the slow part (ethtool counters, PFC and trust) every 30 s. Units without RDMA devices are skipped and re-checked every 10 minutes. The API exposes it as `metrics.roce` in `/api/sparks/:id/metrics`.

Two health findings use it: **a RoCE link went down** (a port that was up, critical) and **RoCE is dropping packets** (RDMA loss counters such as `out_of_buffer` or sequence errors rising for three samples in a row, or port discards / CRC errors rising on two consecutive 30 s readings; a single dropped frame is not reported). Ports that never came up are not reported. Switch telemetry (PFC and ECN statistics on the switch side) is not part of this; it could be added later as an optional integration.

---

## Windows PCs

A Windows PC with an NVIDIA GPU can be monitored with no agent. On the PC:

1. Install the **OpenSSH Server** (Settings → Apps → Optional features → *OpenSSH Server*), then start it and set it to start automatically: `Start-Service sshd; Set-Service sshd -StartupType Automatic` in an elevated PowerShell. Windows Firewall opens port 22 for it.
2. Make sure the NVIDIA driver is installed (`nvidia-smi` works in a terminal).
3. In sparkDash choose **Add Spark / GPU host**, set **Unit type** to *Windows PC with an NVIDIA GPU*, enter the PC's LAN IP and your Windows user (a password works; for key login see the administrator-account note below).

**Keys work everywhere but this PC?** For a Windows user in the *Administrators* group, `sshd` ignores `~\.ssh\authorized_keys` and reads `C:\ProgramData\ssh\administrators_authorized_keys` instead, and refuses it unless only Administrators and SYSTEM can write to it. Put the public key in that file, then fix its permissions in an elevated PowerShell:

```powershell
icacls.exe "$env:ProgramData\ssh\administrators_authorized_keys" /inheritance:r /grant "Administrators:F" /grant "SYSTEM:F"
```

A refused key on a Windows unit also shows this hint in the connection error.

sparkDash then runs two short PowerShell scripts over SSH per poll (no files are installed): GPU temperature, utilisation, power, VRAM and GPU processes from `nvidia-smi`, and RAM, uptime, CPU load, disks (fixed drives) and network adapters from Windows' CIM classes. Shutdown works (`shutdown.exe /s`). Not available on Windows: CPU temperature (Windows exposes no unprivileged sensor; CPU power is estimated from load), automatic Wake-on-LAN MAC detection (enter the MAC by hand), Hermes and Tailnet checks, model launchers, and the kernel Xid/OOM events. LLM servers on the PC (llama.cpp, Ollama, LM Studio, vLLM) are probed over HTTP as for any unit; if the server only listens on `127.0.0.1`, benchmarks fall back to an SSH tunnel.

**AMD-iGPU Windows hosts (Strix Halo)**: if `nvidia-smi` is not installed, the same GPU script falls back to Windows' GPU Engine / GPU Adapter Memory performance counters — utilization is the busiest engine per physical adapter (three samples per poll, clamped to 100%), and memory comes from the adapter counters, with the total read from the driver's reported memory size (on a unified-memory iGPU this is the BIOS carve-out, e.g. 96 GB on a 128 GB Strix Halo; a MemTotal/MemAvailable fallback applies when the driver does not report it). The unit type is the same *Windows* option. An NVIDIA PC whose `nvidia-smi` is installed but failing does not use this fallback; it reports the failure instead. AMD publishes no GPU temperature or power to Windows, so those read 0. The model launchers are bash/systemd and stay unsupported on Windows; HTTP probing and benchmarks work as usual.

---

## LLM server presets & backend auto-detection

### Server presets

LM Studio (:1234), Ollama (:11434) and TensorFold (:8888) are one-click presets in three places — the Add wizard's LLM-ports step, the unit page's "+ Add LLM port" panel (clicking a preset adds the port immediately), and each LLM panel's settings (the preset fills the port field, then Save). Presets fill a **port number only**: they do not install, start or switch anything, and the dashboard will show a port as unavailable until a server actually answers there.

### How each backend is detected

| Engine | Signal |
|--------|--------|
| llama.cpp | `/slots` with per-slot timings |
| vLLM | Prometheus `/metrics` (the default for OpenAI-compatible servers) |
| SGLang | native `/server_info` / `/get_server_info` + `/model_info` |
| LM Studio | `/api/v0/models` — model entries carry a load `state` (a contract no other engine has) |
| Ollama | native `/api/tags` |
| ds4-server | `ds4_*` Prometheus series |
| EXL3 | `/health` with backend + token totals |
| q27 | Prometheus exposition |
| TensorFold | `owned_by` on `/v1/models`, totals from `/health` |
| FreeToken | `/v1/stats` throughput contract |

Detection is **positive-signature first**: some desktop engines (LM Studio) answer HTTP 200 with a JSON body on *every* path, which would trip naive heuristics ("does `/server_info` return JSON?" → SGLang, or the vLLM default). sparkDash probes each engine's unique endpoint and never lets a catch-all-200 server downgrade an identified backend.

### Decode / prefill tok/s on engines without counters

vLLM, SGLang and friends publish cumulative token counters, and the dashboard diffs them per poll. LM Studio and Ollama publish none, so their tok/s tiles stay at 0 in the LLM panel. sparkDash never sends a request of its own to a model server to measure speed (that would load models, keep them resident and add traffic you did not ask for). The decode and prefill benchmarks and the Prompt Showcase measure real speed on demand.

### Multiple LLM ports

Any unit can monitor several LLM servers on different ports at once; each port gets its own panel with independent backend detection, metrics and history. Panels are matched to their port explicitly, so adding, removing or switching ports never shows one port's data on another's panel.

### Engines run in containers (sparkrun, `docker run`)

sparkDash only needs the engine's HTTP port; it does not care how the engine was started. For a vLLM (or any supported engine) that a container tool such as sparkrun publishes on the unit:

- Add the unit, then add that port in the Add wizard's LLM-ports step or the unit page's **+ Add LLM port** panel. The probe, live tok/s, KV-cache tiles (vLLM Prometheus `/metrics`) and the decode/prefill/quality benchmarks run over HTTP; nothing is installed in the engine's container.
- The dashboard container uses `network_mode: host`, so a port published on the unit's loopback or LAN address is reachable as `127.0.0.1` (this unit) or its LAN IP (a remote unit). Benchmarks on a remote unit whose server only listens on loopback fall back to an SSH tunnel.
- A Docker-based start script that detaches (`docker run -d`) shows **Stopped** in the Models panel because the script's process group is gone, while the port probe shows **Serving**. Treat the port status as the source of truth (the **Output** button re-opens the script's output while it runs).
- Provisioning models (pulling weights, choosing a recipe) stays with your own tool; the Models panel only runs the `start.sh` / `stop.sh` you register.

## Quick start

```bash
git clone https://github.com/MiaAI-Lab/sparkDash.git
cd sparkDash

# Production (Docker; loopback-only by default)
docker compose up --build -d

# Or development (host, with hot reload)
npm install
npm run dev
```

- **Docker**: open **http://127.0.0.1:5555** on the host (arm64 image, auto-restart, host mounts for GPU/metrics access). The compose file serves the frontend from the host's `./dist` when it has been built (`npm run build`) and from the copy baked into the image otherwise, so a fresh clone works; a host build that is older than the image's copy is ignored, so after `git pull` one `docker compose up --build -d` is enough
- **Dev**: Vite on **http://localhost:5173** (proxies API/WS to Express)

For another computer, keep the server on loopback and use an SSH tunnel:

```bash
ssh -N -L 5555:127.0.0.1:5555 user@sparkdash-host
```

Then open `http://127.0.0.1:5555` on that computer. For shared access, use an authenticated TLS reverse proxy, Tailscale Serve, or set `BIND_HOST=0.0.0.0` **and** `SPARKDASH_TOKEN`. A direct LAN bind without a token is **open by default**: anyone who can reach the port can change settings and power units off, and the header shows an **Open access** warning. Set `SPARKDASH_TOKEN` to require a token, or `SPARKDASH_ALLOW_OPEN_REMOTE=0` to make a tokenless LAN bind refuse to start.

When the server has `SPARKDASH_TOKEN` set, the dashboard asks for it: the first request or live-telemetry connection the server turns away opens an **Access token** dialog. Enter the token once; it is checked against the server, stored in this browser only, and the live connection reconnects with it — no reload, no devtools. **Settings → Access token** shows whether one is stored and lets you change or clear it.

For development with Docker (source-mounted, HMR):
```bash
docker compose -f docker-compose.dev.yml up --build
```

**Published image (no local build):** each release tag (`v*`) is built for arm64 and amd64 and pushed to `ghcr.io/miaai-lab/sparkdash`. `docker-compose.image.yml` runs that image with the same host mounts and environment as `docker-compose.yml`; it is the only file the host needs, and `./config` next to it holds the runtime state:

```bash
curl -fsSLO https://raw.githubusercontent.com/MiaAI-Lab/sparkDash/main/docker-compose.image.yml
docker compose -f docker-compose.image.yml up -d
```

It follows `latest` unless `SPARKDASH_VERSION` names a release (`SPARKDASH_VERSION=<version>`); update with `docker compose -f docker-compose.image.yml pull` and `up -d`. The frontend history window (`VITE_HISTORY_HOURS`) is fixed at the default 8 hours in the published image; build locally to change it.

**Remote units + SSH keys (Docker):** SSH is executed *inside* the container on the sparkDash host (typically the head DGX). Configured LAN IPs are from **that** host’s point of view, not your laptop. OpenSSH looks for keys under `/root/.ssh` in the container — the host user’s `~/.ssh` is not used unless you bind-mount it. Uncomment this volume in `docker-compose.yml` (and recreate the container):

```yaml
- ${HOME}/.ssh/id_ed25519:/root/.ssh/id_ed25519:ro
```

If the key file has a non-default name (e.g. `id_ed25519_shared`), mount it **as** `id_ed25519`, or set `SSH_IDENTITY_FILE` to the path inside the container. Keep the file mode `600`. The unit that runs sparkDash itself should be added with **This host (local collectors — no SSH for metrics)**.

---

### Running it on a machine that is not a Spark

The dashboard does not have to run on a Spark. Every unit, including the machine running the dashboard, can be monitored over SSH, so you can run sparkDash on any Linux box (an x86 mini PC, a NAS, a VM) and add all your Sparks as remote units.

- The shipped `Dockerfile` and `docker-compose.yml` target arm64 DGX Spark hosts (platform pin and `aarch64` NVIDIA library mounts). On x86 either edit those lines or run it bare: `npm install && npm run build && npm start` (Node 22).
- Add each Spark with **Add Spark** using its LAN IP and an SSH user. The machine running the dashboard does not need a GPU; skip adding it as a unit.
- An x86 Docker build is untested. Reports and PRs are welcome.

## Architecture

Design principle: **one Spark model, N instances**. Every unit is a record in `config/sparks.json` with a `kind` field (`spark` or `host`). The same `SparkMonitor`, `SystemCollector`, and `LlmProbe` code runs for all of them. Adding a unit is a config change, not a code change.

```txt
┌────────────────────── Docker container (sparkDash) ────────────────────────┐
│  Express (server/)                                                         │
│  ├─ config/sparks.json        Spark registry (API read/write)              │
│  ├─ SparkRegistry             load/persist Sparks; change events           │
│  ├─ SparkMonitor (per Spark)  collector + LLM probe + rate baselines       │
│  │   ├─ SystemCollector       local sysfs/proc OR remote SSH               │
│  │   └─ LlmProbe              HTTP to host:LLM_PORT, backend autodetect    │
│  ├─ REST /api/*                                                            │
│  └─ WebSocket /ws             snapshot stream to browsers                  │
│  React SPA (src/)  — Overview + per-Spark pages, themes, dialogs           │
└────────────────────────────────────────────────────────────────────────────┘
         │ SSH (key or sshpass)                    │ HTTP :8888
         ▼                                         ▼
    remote Spark(s)                         each Spark’s LLM server
```

### Data flow

```txt
Browser  ←→  WebSocket /ws   ←→  SparkMonitor.snapshot()  ←→  collectors
Browser  ←→  REST /api/*     ←→  SparkRegistry + SparkMonitor
```

Poll loops run in the background (even with no clients) so rate metrics — tokens/s, network bytes/s, disk I/O — stay correct.

---

## Tech stack

| Layer | Stack |
|-------|--------|
| Frontend | React 19, TypeScript, Vite 8, Tailwind CSS v4 |
| Backend | Node.js (ESM), Express 5, `ws` |
| Platform | The dashboard targets ARM64 — DGX Spark GB10 (Neoverse V2). Units it monitors can be Linux, macOS or Windows machines over SSH |
| Deploy | Docker multi-stage (arm64), Compose |
| Secrets | AES-256-GCM SSH password store |
| Ports | **5555** dashboard/API; **5173** Vite (dev only) |

---

## Repository layout

```txt
sparkDash/
├── src/                 React + TypeScript SPA
│   ├── api/             REST client + shared types
│   ├── components/      Overview, Spark pages, dialogs, UI primitives
│   ├── hooks/           WebSocket snapshot, routing
│   └── theme / CSS      Tailwind v4 + four themes
├── server/              Express + WebSocket (plain JS ESM)
│   ├── sparks/          SparkRegistry, SparkMonitor
│   ├── collectors/      SystemCollector, LlmProbe, ssh, RoCE, Windows (PowerShell)
│   ├── health/          HealthEvaluator (findings)
│   ├── energy/          Fleet energy tracker + monthly archive
│   ├── llmtokens/       Token ledger · events/ Activity log · metrics/ GPU history
│   ├── llmlaunch/       Model launchers · tooleval/ Tool Eval Bench
│   ├── secretsStore.js  Encrypted password persistence
│   └── validate.js      Host/user validation (SSRF-minded)
├── config/              Runtime state (volume; secrets gitignored)
├── assets/              Logo (bolt.svg) and legacy media
├── .github/             README screenshot and Showcase video; image publish workflow
├── Dockerfile           Production multi-stage arm64
├── docker-compose.yml   Production
├── docker-compose.image.yml  Production from the published image
├── docker-compose.dev.yml
└── deploy.sh            Rebuild / recreate helpers
```

---

## REST API

| Method | Path | Purpose |
|--------|------|---------|
| GET | `/api/sparks` | List Sparks (passwords redacted) |
| POST | `/api/sparks` | Add Spark and start its monitor |
| PATCH | `/api/sparks/:id` | Update Spark (hot-swap config) |
| DELETE | `/api/sparks/:id` | Remove Spark and drain monitor |
| PUT | `/api/sparks/order` | Persist tab order |
| GET | `/api/sparks/:id/metrics` | One-shot metrics snapshot |
| GET | `/api/fleet-energy` | Estimated fleet watts, rolling energy, coverage, and Wh/output-token |
| GET | `/api/fleet-energy/history` | Hourly / daily energy buckets for the Fleet energy page |
| GET/DELETE | `/api/fleet-energy/monthly` | Permanent per-UTC-month energy totals; `DELETE` needs `?month=YYYY-MM` or `?all=true&confirm=delete-all-history` |
| GET | `/api/llm-token-totals` · `/api/llm-token-totals/history` | Cumulative LLM tokens by model, and the hourly / daily history behind the Token totals page |
| GET | `/api/sparks/:id/gpu-history` | Last hours of GPU utilization, temperature and power % (`windowMs`, up to 8 h; parallel arrays) |
| GET/POST | `/api/restart` | `GET` says whether the server can restart itself (Docker with a restart policy); `POST` shuts down cleanly so Docker starts it again |
| DELETE | `/api/events` · `/api/fleet-energy` | Clear Activity or Fleet energy: everything, or only entries older than `?olderThanMs=` |
| DELETE | `/api/llm-token-totals` | Reset Token totals (all Sparks, or one with `?sparkId=`) |
| GET | `/api/health` · `/api/auth/status` | Bind / auth mode; whether a token is required and accepted |
| POST | `/api/sparks/:id/shutdown` · `/wake` · `/api/sparks/shutdown-all` · `/wake-all` | Power controls (see [Power controls](#power-controls-shutdown--wake-on-lan)) |
| POST | `/api/sparks/:id/refresh/:domain` | Re-poll one domain now (`storage` or `llm` only) |
| GET/POST/DELETE | `/api/sparks/:id/llm/showcase[/:sessionId]` | Prompt Showcase sessions |
| GET | `/api/events` | Fleet event log (`limit`, `sparkId`, `sinceId`, `beforeId`; 2000 kept) |
| GET/POST/PUT/DELETE | `/api/sparks/:id/llm-launchers[/:lid]` | Registered start.sh / stop.sh model launchers; `POST …/:lid/start` and `…/stop` run them, `GET …/jobs/:jobId` streams the shell output |
| GET/POST/DELETE | `/api/sparks/:id/tool-eval/…` | Tool Eval Bench on a Spark: `status`, `install`, `preview`, `probe`, `runs` (start / list / stream / attach / stop / refresh / `result` / delete); `GET /api/tool-eval/spec` serves the option spec |
| POST | `/api/sparks/test` | Ephemeral SSH + LLM (+ Comfy if enabled) test (no persist) |
| POST | `/api/sparks/:id/test` | Connectivity test (can save password) |
| POST | `/api/sparks/:id/comfy/cancel` | Cancel ComfyUI job by `promptId` |
| PUT | `/api/sparks/:id/password` | Save SSH password (works offline) |
| PUT | `/api/sparks/:id/disabled-devices` | Hide storage devices (hot) |
| PUT | `/api/sparks/:id/disabled-interfaces` | Hide network interfaces (hot) |
| PUT | `/api/sparks/:id/llm-ports` | Replace all LLM ports (hot) |
| POST | `/api/sparks/:id/llm-ports` | Add an LLM port (hot) |
| DELETE | `/api/sparks/:id/llm-ports/:port` | Remove an LLM port (hot) |
| PUT | `/api/sparks/:id/llm-port` | LLM port — backward-compat (hot) |
| GET | `/api/sparks/:id/llm/daily` | Daily busy decode/prefill tok/s (`port`, `days`) |
| POST | `/api/sparks/:id/llm/bench` | Start decode benchmark (202); poll/cancel/clear on the same path |
| POST | `/api/sparks/:id/llm/prefill-bench` | Start prefill + TTFT context sweep (202); poll/cancel/clear on the same path |
| POST | `/api/sparks/:id/llm/quality-bench` | Start quality suite (202); `GET` lists active / last / history summaries, `GET :benchId` returns a full run, `DELETE :benchId` cancels, `DELETE` clears history |
| GET | `/metrics` | Prometheus exposition (opt-in, see [Prometheus](#prometheus)) |
| GET | `/api/settings` | Global settings |
| PUT | `/api/settings` | Update global settings |
| WS | `/ws` | Real-time metrics stream |

There is no application authentication on the HTTP/WebSocket API. sparkDash therefore binds to loopback and refuses direct LAN binding. Use an SSH tunnel, authenticated TLS reverse proxy, or Tailscale Serve; see [Remote access](./docs/REMOTE-ACCESS.md).

`/api/fleet-energy` samples the configured fleet independently every two seconds. It estimates
each node as GPU board draw + a CPU utilization model (5.2–65 W) + 23 W of memory/network/base
overhead, clamped to the DGX Spark power envelope. Current and hourly fleet watts require fresh,
simultaneous telemetry from every node; coverage fields make gaps explicit. Minute buckets are
persisted at mode `0600` for rolling 24-hour and 31-day windows. Wh/output-token is computed from one observation per available LLM endpoint on each `head` or
`standalone` node (workers are skipped, since they front their head's engine).
These values are estimates, not wall-meter measurements. Restart sparkDash after changing fleet
membership so the persisted series has one stable node set.

Finished minutes are also rolled up once into a permanent monthly archive
(`config/fleet-energy-monthly.json`, read through `/api/fleet-energy/monthly`): per UTC month and
node watt-hours and coverage, the fleet total, output tokens and Wh/output-token. A minute is added
10 seconds after it ends, a stored watermark guarantees it is never counted twice, and the first
start backfills from the existing 31-day file. The archive is not touched by **Reset…**, restarts or
fleet membership changes (the previous scope's file is rolled up, after the same plausibility checks as a normal load, before it is set aside). Only
`DELETE /api/fleet-energy/monthly` removes months (all of them only with `confirm=delete-all-history`). A clock that jumps ahead cannot hide later
minutes: nothing is folded, and no watermark is kept, beyond the clock plus one day (a stored watermark
beyond that is clamped back on load, totals unchanged).

---

## Configuration

### Global settings (UI or API)

Gear icon in the header, or `GET`/`PUT` `/api/settings`:

| Setting | Default | Description |
|---------|---------|-------------|
| Refresh rate (`pollIntervalMs`) | 2000 ms | How often every unit is polled for GPU, CPU/RAM, network, memory bandwidth, LLM and ComfyUI metrics, and how often the dashboard is pushed an update (the dialog offers 1 s to 10 s; the server accepts 0.5 s to 60 s). Remote units are polled over SSH, so 1 s costs the most. Memory bandwidth (`nvidia-smi dmon`, which blocks ~1 s) never goes below 2 s. Storage, liveness, Tailnet, Hermes and the `NV_ERR` scan keep their own cadences. A `POLL_INTERVAL_*` env var, when set, pins its domain instead |
| Default LLM port | 8888 | Default for new Sparks |
| Hide offline Sparks | false | Hide offline Sparks on Overview |
| Hide worker nodes | false | Hide Worker-role Sparks from Overview and the tab bar |
| Temperature unit | Celsius | Display GPU temperature in °C or °F |
| Time zone | Browser | Hour and day boundaries and labels on the Tokens, Energy and Activity charts (IANA name such as `Europe/Paris`; empty follows each browser). Daily token buckets and monthly energy stay UTC |
| Benchmark share image | true | Decode/prefill **Copy results** becomes a split button: the label copies the text summary, the caret offers **Copy as text** / **Copy as image** on hover or click. Turn it off to keep the plain button. The image copies where the page has an image clipboard (HTTPS or localhost); over plain http on a LAN IP the card downloads instead |
| Detailed VRAM breakdown | true | The VRAM bar on the Overview cards and the GPU panel is split by what holds the memory — LLM engine (largest GPU process while an endpoint is serving), system/CPU (GB10 unified pool), other GPU use — over a free track, and turns amber/red on low free memory (GB10: under 8 / 4 GB; discrete GPU: under 2 / 1 GB) rather than on a high percentage. Hover or focus for the breakdown, including the engine's KV fill where the backend reports it. Turn it off for the single percentage bar |
| Show Fleet Energy | true | Overview card with rolling fleet power estimates (the Fleet energy page is always in the sidebar) |
| Electricity price / currency | not set / `$` | Per kWh, for the estimated cost on the Fleet energy page; empty hides cost |
| Show LLM Token Totals | true | Overview card with cumulative tokens per model |
| Show Recent activity | true | Overview card with the latest health and kernel events (the Activity page is always in the sidebar) |
| Show Fleet decode / Fleet power / Memory in use / Tokens today tile | true | Each of the four Overview tiles can be hidden on its own; the rest share the row (Tokens today also covers the Online count that replaces it when no tokens are tracked) |
| Show active fleet exceptions | false | Overview strip for offline units, throttling, disk and LLM alerts |
| Show search and status filters | false | Overview search field and status dropdown |
| Save benchmark debug traces | false | Store prompts, HTTP ids and GPU samples in bench history (larger files) |
| Prometheus metrics | false | Serve `GET /metrics` for Prometheus / Grafana (see [Prometheus](#prometheus)); `404` while off |

### Environment variables

Copy `.env.example` to `.env` if needed:

| Variable | Default | Description |
|----------|---------|-------------|
| `BIND_HOST` | `127.0.0.1` | HTTP and WebSocket listen address. A non-loopback bind without `SPARKDASH_TOKEN` is open to anyone who can reach it (see `SPARKDASH_ALLOW_OPEN_REMOTE`). |
| `SPARKDASH_TOKEN` | _(empty)_ | Bearer token. When set, it is required for every mutation and WebSocket connection, and for REST reads on a non-loopback bind. The browser prompts for it when needed (Settings → Access token to change it). |
| `SPARKDASH_ALLOW_OPEN_REMOTE` | `1` | Unset, empty, or `1`: a non-loopback bind without `SPARKDASH_TOKEN` stays open. `0`: refuse to start without a token (fail closed). |
| `SPARKDASH_ALLOWED_HOSTS` | _(empty)_ | Only for a reverse proxy on a custom domain: comma-separated names a loopback bind should also answer to. `localhost`, IP addresses, this machine's hostname and its Tailscale name work without it. |
| `PORT` | `5555` | HTTP + WebSocket listen port |
| `LLM_PORT` | `8888` | Default LLM probe port |
| `COMFY_PORT` | `8188` | Default ComfyUI probe port |
| `POLL_INTERVAL_GPU` | _(setting)_ | GPU poll (ms). Unset: follows **Settings → Refresh rate**; set: pins GPU polling regardless of the setting |
| `POLL_INTERVAL_COMFY` | _(setting)_ | ComfyUI probe poll (ms); same rule |
| `POLL_INTERVAL_CPU` | _(setting)_ | CPU / RAM poll (ms); same rule |
| `POLL_INTERVAL_NETWORK` | _(setting)_ | Network poll (ms); same rule |
| `POLL_INTERVAL_STORAGE` | `5000` | Storage poll (ms): disk usage and the read/write speed on the Storage panel |
| `POLL_INTERVAL_LLM` | _(setting)_ | LLM probe poll (ms); same rule |
| `POLL_INTERVAL_BANDWIDTH` | _(setting, ≥ 2000)_ | Memory bandwidth / dmon poll (ms); same rule, and when it follows the setting it never drops below 2000 |
| `POLL_INTERVAL_HERMES` | `600000` | Hermes Agent update check poll (ms) |
| `POLL_INTERVAL_TAILSCALE` | `30000` | Tailnet probe poll (ms) |
| `TAILSCALE_PROBE_TIMEOUT_MS` | `8000` | Timeout for `tailscale status --json` (ms) |
| `POLL_INTERVAL_NVERR` | `60000` | Kernel journal scan for NVRM `NV_ERR_NO_MEMORY` (ms) |
| `HERMES_UPDATE_TIMEOUT_MS` | `600000` | Hard timeout for running `hermes update` over SSH (ms) |
| `POLL_INTERVAL_LIVENESS` | `5000` | Online/SSH liveness check (ms) |
| `SPARKDASH_SECRETS_KEY` | _(auto)_ | Passphrase or 64-char hex for secret encryption |
| `HOST_PROC_PATH` | `/host/proc` | Host proc mount inside container |
| `HOST_SYS_PATH` | `/host/sys` | Host sys mount |
| `HOST_ROOT_PATH` | `/host/root` | Host root mount |
| `SSH_IDENTITY_FILE` | _(unset)_ | Path **inside the process** to a private key (`ssh -i`). Use when the bind-mount is not a default OpenSSH name. |
| `SSH_CONTROL_PERSIST_SECONDS` | `60` | Idle SSH transport persistence in seconds, capped at `3600`. Set to `0` to disable multiplexing. |
| `FLEET_ENERGY_JSON_PATH` | `config/fleet-energy.json` | Rolling fleet-energy persistence path |
| `FLEET_ENERGY_MONTHLY_JSON_PATH` | `config/fleet-energy-monthly.json` | Permanent monthly energy archive |

For compatibility, `SSH_CONTROL_PERSIST` is accepted as a seconds-based fallback when `SSH_CONTROL_PERSIST_SECONDS` is unset. The existing `SSH_MULTIPLEX=0` switch also disables reuse. SSH tunnels always use an independent connection.

> The listener and both Compose files default to `127.0.0.1`. Existing Docker users who opened
> `http://<host-ip>:5555` must migrate to an SSH tunnel, authenticated reverse proxy, Tailscale
> Serve, or `BIND_HOST=0.0.0.0 SPARKDASH_TOKEN=...` (without the token a `0.0.0.0` bind is open to the
> network unless `SPARKDASH_ALLOW_OPEN_REMOTE=0`). Recovery:
> `BIND_HOST=127.0.0.1 docker compose up -d --force-recreate`.

### Adding a unit

1. Click **+** next to *Sparks* in the sidebar (or **Add Spark / GPU host** in the Sparks menu on a phone).
2. Choose **Unit type**:
   - **NVIDIA DGX Spark** — the default; hardware summary shows DGX Spark specs and the CX7 IP field is available.
   - **Dedicated GPU host** — any Linux machine with an NVIDIA GPU. It is monitored exactly like a Spark (SSH + `nvidia-smi`) but is **not** reported as a DGX Spark: the header shows a detected hardware summary (GPU model, CPU, RAM) instead of fixed GB10 specs, and the page shows separate **RAM** and **VRAM** panels (VRAM from `nvidia-smi`, RAM from system memory). On the unit page, RAM → Network → Storage stack in the right column with GPU filling the left column. A host with **more than one GPU** needs nothing extra: every card `nvidia-smi` lists is collected, the header names them all, the GPU panel shows a block per card, and `metrics.gpu` stays the aggregate (hottest / busiest card, summed power and VRAM) with the per-card detail under `gpu.gpus[]`.
   - **Windows PC with an NVIDIA GPU** — a Windows machine with the OpenSSH Server and the NVIDIA driver; see [Windows PCs](#windows-pcs). Always a remote unit; no model launchers, CPU temperature or Hermes / Tailnet checks.
3. Set **Name** and choose whether this is **This host**. Local units do not require a LAN IP or SSH; their optional LAN IP enables browser links. On a dedicated GPU host it also directs Wake-on-LAN. Remote units require a LAN IP/host, SSH user, and key or password. Key auth in Docker needs a key mounted into the container (see Quick start).
   - **LLM host** (optional, `llmHost` in the unit's config): when the model API listens on one specific address of a multi-interface Spark, pin it here; probes and benchmarks then use it instead of the LAN IP (SSH keeps using the SSH host). An unreachable `llmHost` is never tunnelled to the SSH host's loopback.
4. **Test** shows pass/fail/skipped for host collectors/SSH and each enabled service (LLM, ComfyUI, Hermes Agent, Tailnet). Every enabled capability must pass; disable an unavailable optional service before saving if it should not be monitored.
5. Save — a tab appears and metrics start streaming.

### Loaded models in private containers

Overview lists all models returned by the configured, reachable LLM endpoints.
For remote Linux units with LLM monitoring on (never workers), it also checks running Docker containers whose image name
contains `vllm`, every 30 seconds over the existing SSH connection. It reads
`/v1/models` on each container's private network address and configured `--port`
(default 8000), without publishing ports or sending inference requests.

This optional discovery requires Python 3 and existing permission to read Docker
state on the remote unit. At most 20 containers and 50 models are read, within a few
seconds; a failed check keeps the previous list for about 90 seconds before clearing it.
Local, macOS and Windows units are skipped. Additional models are
labelled **Loaded**. Throughput, benchmarks and token totals still apply to the
configured LLM endpoints; discovery does not add a monitored endpoint. Offline
units do not display stale discovered models.

### Power controls (shutdown / Wake-on-LAN)

- **Shutdown** (per Spark or **Shutdown All** on Overview) runs the host helper
  `/usr/local/bin/spark-shutdown` with passwordless sudo. The helper contract is
  two invocations:

| Invocation | Expected behaviour |
  |------------|--------------------|
  | `spark-shutdown` | Schedule the graceful shutdown |
  | `spark-shutdown --check` | Print an acknowledgement, exit 0, change nothing |

`--check` is what proves authorization before anything is scheduled, so a
  sudoers rule scoped to the helper is enough:

```
  sparky ALL=(root) NOPASSWD: /usr/local/bin/spark-shutdown
  ```

A helper without `--check` still works when sudo is granted more broadly (the
  authorization probe falls back to `sudo -n true`), but a rule limited to the
  helper path needs `--check` support.
- On a **local unit**, the helper runs on the Spark itself. When the dashboard
  is in Docker that means the invocation first enters the host mount namespace
  (`nsenter --mount=/host/proc/1/ns/mnt -- sudo -n …`), using the same
  `HOST_PROC_PATH` mount and `privileged: true` the collectors already need. A
  bare-host install calls `sudo` directly. The helper always resolves against
  the **host** filesystem, so it does not need to exist inside the container.
- **Wake** / **Wake All** send a UDP magic packet (port 9) to **dedicated GPU hosts** whose NIC and firmware support Wake-on-LAN. The MAC is taken from the **enP7s7** interface while the host is online (persisted as `detectedMacAddress`), or from a **MAC override** in Edit Spark. Broadcast is derived as `/24` from LAN IP, or `255.255.255.255` if LAN IP is missing. **DGX Spark does not wake this way** — the GB10 onboard NIC has no Wake-on-LAN — so the Wake control is not offered on Spark units.
- On a **Windows PC** shutdown runs `shutdown.exe /s /t 5` over SSH; Wake-on-LAN needs a MAC address entered by hand.
- Batch shutdown only targets **online** Sparks; offline nodes are skipped.
- Power APIs are mutations: with `SPARKDASH_TOKEN` set they require it; without it they are open on loopback (local trust) **and** on a remote bind, unless `SPARKDASH_ALLOW_OPEN_REMOTE=0` makes that bind fail closed.

### Themes

Choose a theme in **Settings → Appearance** (or the sun / moon button in the sidebar):

| Theme | Notes |
|-------|--------|
| **Dark** (default) | Neutral grays, true black base, muted amber accent |
| **Light** | Warm paper whites |
| **White** | Cool neutral whites |
| **OLED** | True black for OLED panels |

Choice is stored in `localStorage`.

---

## Security

- **SSH passwords** are not stored in `sparks.json` and are never returned by the API.
- Passwords are encrypted with **AES-256-GCM** in `config/sparks-secrets.json` (survives restarts).
- Encryption key: `config/.secrets-key` (auto-generated) or `SPARKDASH_SECRETS_KEY`. **Do not delete the key file** or encrypted secrets become unreadable.
- **Target validation** rejects clearly unsafe IPv4 targets (link-local `169.254.0.0/16`, `0.0.0.0/8`, multicast/reserved ≥ 224). Private, loopback, and public addresses are allowed so LAN and remote Sparks work.
- SSH and HTTP probes use short timeouts (about 5 s SSH connect, 3 s HTTP) so a hung host cannot stall the poll loop.
- Prefer **SSH keys** over passwords. In Docker, mount the private key into `/root/.ssh` (see Quick start); passwords are the only SSH secret the app stores itself.
- Loopback installs remain local-trust. A remote bind (`BIND_HOST` not loopback) **without** `SPARKDASH_TOKEN` is open by default: anyone who can reach the port can read telemetry, change settings and power units off, and the header shows an **Open access** warning (dismissible per browser). Set `SPARKDASH_TOKEN` to require a bearer token for mutations and remote telemetry/WebSocket, and `SPARKDASH_ALLOW_OPEN_REMOTE=0` to refuse to start a remote bind without one. `GET /api/health` reports which applies as `authMode`: `loopback-open`, `bearer`, `open-remote`, or `required-missing`.
- One-off remote benchmark hosts must be listed in `SPARKDASH_BENCH_HOSTS`.
- Tested operator capacity for this remediation: **12 units**.

---

## Scripts

| Command | Purpose |
|---------|---------|
| `npm run dev` | Vite (5173) + Express (5555) together |
| `npm run dev:server` | Express only (`node --watch`) |
| `npm run dev:client` | Vite only |
| `npm run build` | Production frontend → `dist/` |
| `npm run build:watch` | Rebuild `dist/` on every save |
| `npm run typecheck` | `tsc --noEmit` |
| `npm test` | Server tests (`node --test`) and frontend tests (`vitest`); also `npm run test:server` / `test:frontend` |
| `npm start` | Production server (`node server/index.js`) |
| `npm run docker:up` | `docker compose up -d` |
| `npm run docker:prod` | Same as `docker:up` |
| `npm run docker:rebuild` | `docker compose up --build -d` |
| `npm run docker:dev` | Dev Compose |
| `npm run docker:dev:build` | Dev Compose with rebuild |
| `./deploy.sh` | Recreate container; `--build`, `--frontend` flags |

---

## How it works

### Local vs remote Sparks

One `SystemCollector` path for both modes. When `spark.isLocal` is true, metrics come from host sysfs/proc and `nvidia-smi` (often via nsenter into the host namespace). Remote Sparks wrap the same commands in a shared `sshExec()` helper (key agent or `sshpass`). The helper reuses an authenticated OpenSSH transport by default so frequent metric polls do not create a new SSH/PAM login lifecycle each time. Set `SSH_CONTROL_PERSIST_SECONDS=0` to disable reuse. For `kind: "host"` units, actual hardware (GPU model, driver version, CPU, RAM) is detected once and cached in place of the static DGX Spark specs, and GPU VRAM comes straight from `nvidia-smi` while system RAM is read from `/proc/meminfo`.

### Remote SSH sessions and host memory

Older sparkDash versions could create hundreds of SSH/PAM login sessions per minute on each remote host. [Issue #73](https://github.com/MiaAI-Lab/sparkDash/issues/73) documents the resulting session churn and observed `polkitd` memory growth. Connection reuse reduces this churn while retaining the collector refresh cadence. After updating, verify that metrics keep advancing and that new SSH authentications/PAM session opens fall after the initial connection; a new SSH client process for each collector command is still expected.

If host memory remains low, compare Linux `MemAvailable` and per-process resident/swap usage. Memory retained by `polkitd` requires separate OS investigation: [polkit PR #653](https://github.com/polkit-org/polkit/pull/653) fixes a reference leak in `NoNewPrivileges` queries. Check whether your distribution's polkit package includes that fix. SSH reuse neither applies the OS patch nor releases memory already retained by another process.

### Health findings

`server/health/HealthEvaluator.js` runs a handful of rules over metrics sparkDash already collects, once per Spark, and reports them as `health[]` on the unit's snapshot. A rule never guesses: a missing metric raises nothing. Findings carry a severity, a detail and a hint, show up as chips on the Overview cards and in full on the unit page, and the monitor writes an Activity event when one appears or clears (never for the first baseline after a restart). The kernel Xid / OOM counters come from one cached `journalctl -k` scan per minute.

### Windows units

A unit with `platform: "windows"` is reached over the same SSH helper, but every command is a PowerShell script sent as `-EncodedCommand` (so it works whatever the OpenSSH server's default shell is). Two scripts per poll cover the GPU (`nvidia-smi`, or the GPU Engine/Adapter Memory performance counters on an AMD iGPU) and the CIM classes for RAM, CPU load, disks and adapters; their output is parsed into the same shapes the Linux collectors return. See [Windows PCs](#windows-pcs).

### Graceful degradation

Collectors catch errors and return zero/default metrics instead of crashing the loop. After sustained liveness failures, a Spark is marked offline; the UI shows stale or empty states rather than hard errors.

### Hot configuration

Name, IP, SSH credentials, LLM port, and device/interface filters update the running `SparkMonitor` without tearing down poll loops or losing rate baselines. Registry writes are atomic (temp file + rename).

### LLM probe

Each configured LLM port gets its own `LlmProbe` instance running in parallel. Probes auto-detect backends:

- **llama.cpp** — `/slots` for live decode rates; model from `/props`
- **ds4-server** (Entrpi/ds4-on-spark) — `/v1/models` (`owned_by: ds4.c`) + Prometheus `ds4_*` token counters for live tok/s
- **EXL3** (ExLlamaV3 `tools/serve_openai.py`) — `/v1/models` (`owned_by: exl3`) or `/health` `{ok, busy}`; live tok/s from `/health` cumulative counters
- **q27** (signalnine/q27 engine) — `/v1/models` (`owned_by: q27`) or Prometheus `q27_*` series; live tok/s from `q27_*_processed` counter diffs (completion-based totals as fallback), exact computed-only prefill with the cached/uncached split doubling as the prefix-cache hit rate, TTFT/E2E/ITL p95 histograms, and constant-0 preemptions (FIFO admission, no wait queue)
- **TensorFold** (ashhart/TensorFold) — `/v1/models` (`owned_by: tensorfold`). No `/metrics`. Live tok/s comes from `/health` when it publishes `prompt_tokens_total`, `completion_tokens_total`, and `prefill_seconds_total` (decode is the completion-counter diff; prefill is prompt tokens ÷ prefill time). The slots tile reads `streams.max` and `requests_running` on CUDA 0.6.0, and `max_batch_size` on MLX. A stock server that only returns `{ok: true}` stays at 0. Decode/prefill benches and the showcase work regardless.
- **FreeToken** — `/v1/models` (`owned_by: FreeToken`), or a `/v1/stats` document with `throughput.decode_tps` / `prefill_tps` and lifetime prompt/completion totals. Those rates are FreeToken's own 5-second window (0 when idle). `requests.ttft_mean_ms` is a mean; `requests.p95_ms` is end-to-end latency, not TTFT p95. Queue length, slot capacity, and preemptions stay unset.
- **vLLM / sglang** — `/v1/models`; sglang via `/server_info` (`last_gen_throughput` when metrics off; `/get_server_info` fallback), vLLM via Prometheus `/metrics` counters (scientific notation supported)

Rates are derived from per-probe cumulative counter diffs (or SGLang sticky throughput while it moves). Multiple ports can be added or removed at runtime without restarting the monitor.

Live probes use the LAN IP on remote units, and fall back to an SSH forward to the unit's loopback after two failed direct probes (the LLM panel then shows a **via SSH** chip). **Decode, prefill and quality benches** try that same HTTP target first; if it is closed they open an SSH local-forward onto the remote’s `127.0.0.1` so loopback-bound servers (ds4 `start.sh` default) can still be measured. The tunnel is torn down when the job finishes or is cancelled.

---

## Optional inference request activity

For shared inference deployments, an optional Overview card groups in-flight requests and recent completions by approved **display labels**. It is disabled by default and makes no source requests until configured. The gateway must already expose sanitized JSON; sparkDash does not intercept inference, handle client API keys or infer their owners.

Set these variables in the **server's environment**, then restart sparkDash:

```dotenv
SPARKDASH_INFERENCE_ACTIVITY_URL=http://127.0.0.1:9001/activity
SPARKDASH_INFERENCE_ACTIVITY_LABELS=service-a,user-1
```

The URL accepts HTTP/S with an explicit loopback or RFC1918 IP literal only; DNS names, public/link-local targets, credentials, queries, fragments and redirects are rejected. Requests carry no source credentials, time out after 5 seconds and read at most 64 KiB. There is no browser-editable URL or API-key mapping. These labels are visible to anyone allowed to view the dashboard: protect the dashboard with its normal authentication or private access, and **never configure a key, token, IP, email or prompt as a label**.

The source must return `application/json` with this shape:

```json
{
  "available": true,
  "generatedAt": "2026-10-11T12:00:00Z",
  "active": [
    {"principal": "service-a", "model": "model-a", "count": 2,
     "oldestAgeMs": 900, "newestAgeMs": 100}
  ],
  "recent": [
    {"principal": "user-1", "model": "model-a", "durationMs": 400, "ageMs": 1000}
  ],
  "overflowed": false
}
```

`generatedAt` is the source sampling time, refreshed by the source on each sample. Both arrays are required. The first 256 active input rows are normalized and grouped by principal/model: duplicate groups sum counts, take the maximum oldest age and minimum newest age. Counts are independently positive safe integers, not limited to 256; an unsafe individual count or combined total makes the sample unavailable. Truncated input is marked incomplete. Snake-case `oldest_age_ms`/`newest_age_ms` are also accepted. Principals not in the configured label list become `unknown`; model labels are optional non-secret identifiers. URL/IP labels and recognizable credential-like model segments become `unknown`, but this is not a general secret detector: the source must never place credentials or user content in metadata. Recent rows may provide an actual HTTP `status`, otherwise no status is invented. At most 24 recent rows are retained in memory, with a 10-minute recent window and bounded non-negative ages/durations. Extra fields—including keys, headers, IPs, request IDs and prompts—are discarded, not persisted. Source bodies, failures and URLs are not logged or returned by the dashboard API.

The source is read at most once per 5 seconds, shared between dashboard viewers. The server judges sample freshness: samples older than 30 seconds, invalid payloads and failed reads show **unknown**, not zero. The browser trusts the server's `available`, without comparing server timestamps to its own clock. A separate 30-second no-response timer hides the last live display if a later browser fetch hangs, and resets on a successful response. A fresh empty `active` array is a measured zero. The read-only `/api/inference/activity` route uses the dashboard's existing GET access policy and `Cache-Control: no-store`: loopback GETs remain open, while a token-protected remote bind requires the dashboard token. Configuration is not writable through that route or Settings. The card is absent when no URL is configured, including when the normal sidebar cards are hidden.

**In flight includes queued, prefill and decoding HTTP requests.** It is not the number of active GPU slots, does not distinguish engine queueing, and does not measure per-key token throughput. This card needs an external sanitized activity source; an unmodified model server without that source remains unchanged.

## Contributing

Contributions are welcome. Conventions:

- **Server**: plain JavaScript ESM
- **Client**: TypeScript + React
- Prefer extending the shared Spark model over per-unit special cases
- Run `npm test` and `npm run typecheck` before a pull request. Tests that start the server must point every config file at a temp directory (`server/__tests__/isolatedEnv.js`)
- Read [AGENTS.md](./AGENTS.md) and [CODEBASE.md](./CODEBASE.md) first

---

## License

Versions up to 1.9.0 were released under the MIT License ([LICENSE-MIT](./LICENSE-MIT)), and those releases stay MIT-licensed.

---

## Acknowledgements

sparkDash is built and maintained by [Mia's AI Lab](https://mia-ai.net/).

**Contributors.** Thank you to everyone who sent code, fixes and ideas:
[@MikeGibbsOnyx](https://github.com/MikeGibbsOnyx), [@Lesilva](https://github.com/Lesilva), [@vincenzopalazzo](https://github.com/vincenzopalazzo), [@Acermax](https://github.com/Acermax), [@danielkuykendall23-boop](https://github.com/danielkuykendall23-boop), [@0xdfi](https://github.com/0xdfi), [@BHCC2025](https://github.com/BHCC2025), [@ayylemao](https://github.com/ayylemao), [@SashaMIT](https://github.com/SashaMIT), [@krunkosaurus](https://github.com/krunkosaurus), [@andrei-dotdna](https://github.com/andrei-dotdna), [@0xWhiteMage](https://github.com/0xWhiteMage), [@kesslerio](https://github.com/kesslerio), [@nkavassalis](https://github.com/nkavassalis), [@Olyno](https://github.com/Olyno), [@willy92wins](https://github.com/willy92wins) and [@saitakarcesme](https://github.com/saitakarcesme), plus Nyx Voss (who wrote the macOS collectors with C. Michael Gibbs, @MikeGibbsOnyx), Liao Shiwu and Philip Eriksson, and to everyone who reported issues. The full list is on the [contributors page](https://github.com/MiaAI-Lab/sparkDash/graphs/contributors).
The [`sparkdash-dgx-cluster`](https://github.com/glanceapp/community-widgets/tree/main/widgets/sparkdash-dgx-cluster) Glance widget was contributed by [@linxichen](https://github.com/linxichen).

**Projects and data sparkDash builds on**
- [tool-eval-bench](https://github.com/SeraphimSerapis/tool-eval-bench) by SeraphimSerapis (MIT) powers **Tool Eval Bench**. It adapts the scenario methodology of ToolCall-15 by stevibe (MIT) and credits the Typed Decisions dataset from the LocalLLaMA organization (Apache 2.0).
- **GSM8K** (OpenAI, MIT) and **MMLU** (Dan Hendrycks, MIT) supply the maths and knowledge questions in the Quality bench. The instruction-following category is inspired by **IFEval** (Zhou et al., Google).
- [Hermes Agent](https://github.com/nousresearch/hermes-agent) by Nous Research can be monitored and updated from sparkDash.
- The health findings were inspired by [spark-doctor](https://github.com/joeynyc/spark-doctor) by joeynyc (MIT). No code from it is used.
- Fonts: [Geist and Geist Mono](https://github.com/vercel/geist-font) by Vercel (SIL Open Font License 1.1). The bolt in the logo follows the "zap" icon from [Feather](https://github.com/feathericons/feather) (MIT).
- Built with [React](https://react.dev/), [Vite](https://vitejs.dev/), [Tailwind CSS](https://tailwindcss.com/), [Express](https://expressjs.com/), [ws](https://github.com/websockets/ws), [undici](https://github.com/nodejs/undici), [dnd kit](https://dndkit.com/) and [dotenv](https://github.com/motdotla/dotenv). Licenses and copyright notices are in [THIRD_PARTY_NOTICES.md](./THIRD_PARTY_NOTICES.md).

Built for the **NVIDIA DGX Spark (GB10)** on ARM64.

Repository info

Default branch: main
Latest version: v2.32.0
Created: 2026-07-20
Last activity: 2026-10-13
License: Apache-2.0
Open issues: 16
Watchers: 7

Contributors

MiaAI-Lab

130 contributions

MikeGibbsOnyx

53 contributions

Lesilva

30 contributions

vincenzopalazzo

18 contributions

danielkuykendall23-boop

8 contributions