Collectd-web turns the raw RRD files Collectd writes to disk into a fast, modern dashboard. Below is a thorough walk-through of each capability — what it does, why it matters, and how it fits into the way teams actually run self-hosted monitoring today.
A clean, responsive interface that surfaces the metrics you actually want to look at.
Collectd-web's dashboard is built around the assumption that you don't want to memorize plugin names or remember the exact path to an RRD file. Hosts are grouped along the left, plugins are grouped sensibly under each host, and a single click renders the graph. The layout responds to mobile widths so you can glance at production from a phone without zooming and panning.
Hide the noise, pin the dashboards your team relies on, and theme the UI to match your stack.
Every team monitors a different slice of their infrastructure. Collectd-web lets you reorder plugin groups, hide plugins you don't care about, and override default time ranges per graph. Custom CSS hooks make it straightforward to brand the interface for an internal status page, and the whole UI ships with a dark mode that doesn't fight your terminal.
A single .deb, a Docker image, or a one-line Python invocation — pick whichever is closest to your workflow.
There's no installer wizard, no agent to push, and no SaaS account to create. The Debian package wires up systemd, the Docker image is a few megabytes, and the source-based runner is a Python script you can launch from a checkout in five seconds. All three paths converge on the same configuration file, so moving between them is painless.
Optimized for low-overhead operation across hundreds of hosts and thousands of metrics.
Collectd already does the heavy lifting on each machine. Collectd-web stays out of the way: it reads RRD files on demand, caches rendered graphs briefly, and never holds open background workers you didn't ask for. Even on modest hardware, the UI stays snappy when you're scrolling through a hundred-host fleet.
Beautiful, accurate graphs straight from the round-robin database files Collectd produces.
Round-robin databases are a perfect fit for time-series monitoring, but the default rrdtool output looks like it's from 1999. Collectd-web renders graphs with a modern color palette, sensible axis labels, configurable time windows (hour / day / week / month / year), and clean PNG output you can drop straight into incident reports.
Organize each host's data in its own subdirectory and browse them all from a single pane of glass.
Point Collectd at a flat directory of per-host folders and Collectd-web will pick them up automatically. New hosts appear without a restart, and the central collector pattern (using Collectd's own Network plugin) means you can monitor an entire fleet from one machine. The /docs section walks through this in detail.
Collectd-web exists for the people who already run their own infrastructure and don't want to ship metrics to a third party. There's no telemetry, no license server, and no upsell. The full source lives on GitHub under a permissive license, and the project ships a small surface area on purpose so you can audit, fork, or contribute without spending a weekend reading code.
If you're evaluating Collectd-web for a homelab, a small business, or a production fleet, the same install works in all three contexts. The differences are operational — backup strategy, retention policy, reverse proxy choice — and those are documented separately under the docs section.