Glances

What is Glances?

Glances shows CPU, memory, disk, network, processes, sensors and containers on a single auto-updating terminal screen. It also runs as a web server or an API, which makes it useful for a quick look or for feeding another tool.

Best for

Instant system overview over SSH

Why choose Glances

Glances is the tool you want when you are already on the machine. It prints CPU, memory, disk, network, sensors, processes and container stats on one screen that refreshes itself, and it flags the highest consumer in each category so you do not have to sort a process list in your head. Because it is a single Python program, installing it over SSH takes one command and it runs anywhere Python runs. The same tool can also run as a web server or expose a REST API, which means the quick look you take interactively can later feed a dashboard or a script without switching products. For a fast diagnosis on a machine you control, it is the shortest path from SSH to answer.

Replaces

• htop
• top
• Netdata

Key features

• Single-screen overview of everything on a host
• Runs in the terminal, browser or as a REST API
• Docker and GPU monitoring
• Alert thresholds with notifications

What to watch out for

It is a viewer, not an alerting or historical system — once you close it, nothing is recorded, so it cannot tell you what happened last night. The web interface exists but is not its strongest mode and is not designed to be exposed publicly without care. Its container monitoring depends on the paths and sockets being available to the process, which is an easy thing to get subtly wrong in a containerised deployment. On a host with very many processes or GPUs, the terminal view can become noisy, and the sensor support varies with hardware and kernel drivers.

How to deploy

• Install from package manager or pip
• Run in web mode for remote viewing
• Export to InfluxDB/Prometheus for history

Getting started

Install it with pip or your distribution's package and run it once interactively to confirm sensors and disks are detected. Docker users should get the official image and mount the Docker socket and host paths correctly, since partial mounts produce a view that looks complete but is missing data. Set the refresh interval and the units you prefer, and enable whichever extra plugins you actually need rather than all of them. If you want the web mode, put it behind a reverse proxy with authentication and do not publish the port directly. Treat the REST API as read-only monitoring data rather than a control surface.

Typical setup

Usually nothing more than a pip install or a container on the machines being inspected, run interactively over SSH when something looks wrong. Where the web mode is used, it sits behind a reverse proxy with authentication and is never published directly. Container deployments mount the Docker socket and host filesystem paths so the view is complete rather than subtly partial. It occupies no persistent storage and holds no history, so it appears in no backup routine — the things it feeds, such as a metrics exporter or a script polling its API, are what get maintained.

Who should look elsewhere

Do not rely on Glances for anything that needs to happen while you are not watching. If you need history, alerting or a fleet-wide view, this is the wrong shape of tool. It is also not a substitute for a real metrics pipeline on a busy server — it samples on demand rather than collecting continuously, so spikes between refreshes are simply missed.

Project health

• GitHub stars: 33,715
• Last code push: 2026-10-01
• Open issues: 125
• Status: actively developed

Figures pulled from the GitHub API and refreshed periodically.