CF-Server-Monitor

Agents report CPU, memory, disk, network and latency to a Worker, which stores the data in D1 and pushes live updates to browsers over WebSockets through a Durable Object. CF-Server-Monitor also alerts on resource use, monthly traffic and server expiry dates.

The agent on each monitored machine only reports metrics to the Worker. It has no WebSSH or remote command channel and can run without root. The default agent, cfsm-agent, is written in Go and also ships as a Docker image. The older Shell and PowerShell agents are still available but no longer maintained. The agent runs on mainstream Linux distributions, Alpine, OpenWrt, Synology DSM, fnOS, FreeBSD, macOS and Windows. The dashboard covers CPU, GPU, memory, disk, network, load and uptime, among other metrics, in bar, ring, table and map views. History charts go back seven days, and you can track and correct monthly traffic. Latency and packet loss are measured against configurable ping nodes. In the admin panel you can add, reorder, import and export servers, pick a theme from the theme store, and set up notifications through Telegram, SMTP email, a custom webhook or other built-in channels. A single Worker serves the Vue dashboard from static assets, receives agent reports on /update, writes them to D1 with monthly table rotation and broadcasts live data through a SQLite-backed Durable Object. One cron trigger runs offline and resource alerts every minute; another handles hourly maintenance such as table rotation and expiry checks. API_SECRET serves as both the agent reporting key and the initial admin password. The maintainer recommends a GitHub fork connected to Workers Builds, or GitHub Actions, over the one-click deploy, because those paths make upstream updates easier.

Built for: Developers and small teams running a handful of VPS or home servers who want low-cost, low-maintenance monitoring

What you can use it for

• Watch CPU, memory, disk and traffic across a small VPS fleet
• Get alerts when a server goes offline or crosses a resource threshold
• Track renewal dates for rented servers
• Measure latency and packet loss from each server to chosen ping targets

Deploy

• One-click · main path

  Branch main

  Deploy to Cloudflare

• CLI

  Branch main

  npm install
  npm run deploy
  npx wrangler secret put API_SECRET

  Copy

Deploy, then add API_SECRET in the dashboard, change the default admin login, and run the generated agent install command on every server.

What it needs to run

• D1 databasesserver-monitor-db
• Durable ObjectsMetricsBroadcaster

Good to know

Each monitored server needs an agent installed. The one-click deploy does not prompt for API_SECRET, so nobody can sign in until you add it. The default ping nodes target the Chinese carriers China Telecom, China Unicom and China Mobile, plus BGP routes. The repository had no license file until September 27, 2026, when the MIT license was added.

Cost on Cloudflare

Where this comes from

• Repositoryhttps://github.com/huilang-me/CF-Server-Monitor
• App websitehttps://demo.huilang.me/
• Documentationhttps://github.com/huilang-me/CF-Server-Monitor/blob/main/README-en.md

An alternative to Netdata Cloud.