All work

05 · Cloud Computing CoE, CIT · 2025–26

SysDash

Real-time system monitoring: a terminal dashboard on PyPI, and a concurrent Go WebSocket backend feeding a React dashboard.

4K+
PyPI downloads
7
releases of sysdash-cli
1 s
default refresh
SysDash terminal dashboard with CPU, memory, disk, network and process panels

Try it live

See how it works.

Launch film

Film
Launch film · SysDash

A launch film made in Remotion. The web dashboard clip is a real recording; the terminal and theme scenes are recreated from the app's behaviour, with illustrative readings.

The real terminal

Real
Real recording · sysdash-cli
sysdash-cli running in a terminal: CPU, memory, disk, network and process panels updating live

Drive it yourself

Interactive replica

SysDash reads your operating system's metrics, which a browser can't access, so this replica uses random data with the real layout, labels, thresholds and keys.

Live demo · SysDash · terminal dashboard
SysDash CLI | Host: arka | Up time: 1:02:03 | 2026-11-07

CPU & Memory

CPU 24.0% ████░░░░░░░░░░░░░░░░

Memory 46.0% (7.4/16.0 GB) █████████░░░░░░░░░░░

0 █░░░░░░░ 20% 1 █░░░░░░░ 20% 2 █░░░░░░░ 20% 3 █░░░░░░░ 20% 4 █░░░░░░░ 20% 5 █░░░░░░░ 20% 6 █░░░░░░░ 20% 7 █░░░░░░░ 20%

Usage Trends

CPU Max: 24.0% Min: 24.0%

Top Processes

PIDNameCPUMem

Disk Usage

/ 168/460GB ██░░░░░░ 36%

/home 311/700GB ███░░░░░ 44%

/boot 0.4/1GB ███░░░░░ 41%

I/O Rates

Read 4.20 MB/s

Write 2.10 MB/s

Network

Current Rates

Upload 52.0 KB/s

Download 480.0 KB/s

Total Transfer

Sent 1.02 GB

Received 6.41 GB

h help1 cpu2 mem3 disk4 net5 prock killq quit

Simulated with random data. The layout, labels, colour thresholds and keys (1–5, ↑↓, k, y/K/n, h, q) mirror the sysdash-cli terminal UI. Click the dashboard or the key buttons, or type.

Under the hood

How it works.

Step by step, from the first input to what you see.

  1. A collector samples the machine

    A Python collector reads CPU, per-core load, memory, swap, disk I/O, network rates and the top processes, and POSTs JSON to the backend.

    • Python
    • psutil
  2. A concurrent Go backend

    gin receives metrics at POST /api/metrics. A WebSocket hub built on Go channels (broadcast, register, unregister) fans each update out to every client, with a per-client send queue and keep-alive pings.

    • Go
    • gin
    • gorilla/websocket
  3. A live React dashboard

    The dashboard subscribes to ws://…/live and falls back to HTTP polling. Chart.js draws CPU, memory, disk and network; alerts fire on high usage.

    • React
    • Chart.js
  4. Or stay in the terminal

    sysdash-cli renders the same metrics with rich: per-core bars, usage trends, disk and network panels and a process list you can kill from (it asks first).

    • rich
    • PyPI

Honest notes

What broke, and what changed

  • Settings saved but changed nothing

    Theme, accent and chart options were stored but never applied to the interface. They now drive CSS variables, so dark and light themes, four accent palettes and the chart options (smooth lines, points, grid, legend) take effect and persist across reloads.

  • A purple accent that wasn't

    The old Purple option actually rendered cyan. It was replaced with four palettes chosen from the reference colours, and saved settings that still say purple migrate to cyan.

Built with

  • Go
  • gin
  • WebSockets
  • React
  • Chart.js
  • Python
  • rich
  • Docker