Quickstart

This page gives a good introduction to how to get started with Glances. Glances offers multiple modes:

  • Standalone (TUI)

  • Server (REST API and Web UI) / Client

  • Fetch

Standalone Mode

If you want to monitor your local machine, open a console/terminal and simply run:

$ glances

Glances should start (press ‘q’ or ‘ESC’ to exit):

_images/screenshot-wide.png

It is also possible to display RAW (Python) stats directly to stdout using:

$ glances --stdout cpu.user,mem.used,load,network.wlp2s0.bytes_recv
cpu.user: 25.0
mem.used: 954605568
load: {'min1': 2.09, 'min5': 1.01, 'min15': 0.72, 'cpucore': 4}
network.wlp2s0.bytes_recv: 13479
cpu.user: 42.6
mem.used: 1033232384
load: {'min1': 2.24, 'min5': 1.06, 'min15': 0.74, 'cpucore': 4}
network.wlp2s0.bytes_recv: 12868
...

or in a CSV format thanks to the stdout-csv option (key not supported in this mode):

$ glances --stdout-csv now,cpu.user,mem.used,load
now.iso,now.custom,cpu.user,mem.used,load.min1,load.min5,load.min15,load.cpucore
2026-10-11T15:12:05+00:00,2026-10-11 15:12:05 UTC,47.4,1062043648,2.22,1.08,0.74,4
...

or as a JSON format thanks to the stdout-json option (attribute not supported in this mode):

$ glances --stdout-json cpu,load
{"cpu": {"user": 50.0, "system": 6.3, "idle": 43.7, "nice": 0.0, "iowait": 0.0, "irq": 0.0, "steal": 0.0, "guest": 0.0, "total": 56.3, "cpucore": 4, ...}, "load": {"min1": 2.22, "min5": 1.08, "min15": 0.74, "cpucore": 4}}
...

Note: --stdout displays one line per stat per refresh, --stdout-json one JSON object per refresh and --stdout-csv one line per refresh.

Use --stop-after <n> to stop after n refreshes.

Server Mode (REST API and Web UI)

_images/screenshot-web.png

If you want to remotely monitor a machine called server, from any device with a web browser, just run the server with the -s option (-w is the same mode, kept from Glances v4):

server$ glances -s -B 0.0.0.0

then, on the client, enter the following URL in your favorite web browser:

http://@server:61208

where @server is the IP address or hostname of the server.

The server runs headless (no TUI). It serves the RESTful API on /api/5 (see Restful/JSON API documentation) and the Web UI on /. Use --disable-webui to only serve the API.

The default bind address is 127.0.0.1: the server only answers on the local machine. Use -B ADDRESS (or bind_address in the [outputs] section of the configuration file) to listen on another interface, 0.0.0.0 for all the IPv4 interfaces. The default TCP port is 61208, it can be changed with -p PORT (or port in [outputs]).

To change the refresh rate of the page, add the period in seconds to the URL. For example, to refresh the page every 10 seconds:

http://@server:61208/?refresh=10

The Glances web interface follows responsive web design principles.

Here’s a screenshot from Chrome on Android:

_images/screenshot-web2.png

Security — DNS rebinding protection

A server bound to a loopback address (the default 127.0.0.1) only accepts requests whose HTTP Host header is localhost, 127.0.0.1 or [::1]. A server bound to another address applies no Host filtering and logs a warning at startup.

Set the accepted Host values with webui_allowed_hosts in the [outputs] section of glances.conf, for example behind a reverse proxy that forwards its own Host or for any deployment on a non-loopback interface:

[outputs]
webui_allowed_hosts=localhost,127.0.0.1,myserver.example.com

Requests whose Host header does not match any of the listed patterns are rejected with 400 Bad Request. Wildcards are supported, for example *.example.com.

Client Mode

You can also display the stats of a remote Glances server in your terminal. Run the server as described above (glances -s -B 0.0.0.0) and, on the client:

client$ glances -c @server

where @server is the IP address or hostname of the server. It is also possible to give @server:port or an http(s):// URL (for a server behind a TLS reverse proxy).

In client mode, you can set the TCP port of the server with -p PORT.

The client only speaks to a Glances v5 server (REST API /api/5); the Glances v4 XML-RPC server is not supported anymore.

In client/server mode, limits are set by the server side.

Central Glances Browser

_images/browser.png

Glances can centralize available Glances servers using the --browser option. The server list can be statically defined via the configuration file (section [serverlist]).

Example:

[serverlist]
# Define columns (comma separated list of <plugin>:<field>:(<key>)) to grab/display
# Default is: system:hr_name,load:min5,cpu:total,mem:percent
# You can also add stats with key, like sensors:value:Ambient
columns=system:hr_name,load:min5,cpu:total,mem:percent,memswap:percent
# Define the static servers list (the port defaults to 61208)
server_1_name=xps
server_1_alias=xps
server_1_port=61208
server_2_name=win
server_2_port=61235
# The name can also be an http(s):// URL
server_3_name=https://glances.example.com

Glances can also detect and display all Glances v5 servers available on your network via the zeroconf protocol (browser extra needed):

To start the TUI Central Glances Browser, use the following option:

client$ glances --browser

When the list is displayed, you can navigate through the Glances servers with up/down keys and open one with ENTER. It is also possible to sort the server using:

  • ‘1’ is normal (do not sort)

  • ‘2’ is using sorting by status (UNKNOWN > OFFLINE > PROTECTED > UNSUPPORTED > ONLINE)

  • ‘3’ is using sorting by status in reverse order (ONLINE > UNSUPPORTED > PROTECTED > OFFLINE > UNKNOWN)

UNSUPPORTED means that something answers on the port, but not the Glances v5 API (a Glances v4 server, for example).

A server discovered on the network is listed as <name> (zeroconf <address>) and opened at that address: anyone on the network can announce under any name, so check the address before typing a password (it is sent over clear HTTP). An announcement that uses the name or alias of a [serverlist] server is ignored.

To start the WebUI Central Glances Browser, use the following option:

client$ glances --browser -s

Open the URL (/browser) and click on the server to display stats. The servers list is also available on the /api/5/serverslist endpoint.

Note

Use --disable-autodiscover to disable the auto-discovery mode.

IPv6

Glances is IPv6 compatible. Just use the -B :: option to bind to all IPv6 addresses.

How do you protect your server with a login/password ?

Generate a password hash with the --set-password option:

$ glances --set-password
Password:
Confirm:

Paste this value into [outputs] password of your glances.conf:

<salt>$<hash>

and paste it in the [outputs] section of the Glances configuration file (~/.config/glances/glances.conf on GNU Linux operating system). By default, the login is glances but you can change it with the username key:

[outputs]
password=<salt>$<hash>
# Optional, default is glances
#username=glances

The REST API and the Web UI then require this login/password (HTTP Basic authentication or a JWT bearer token from /api/5/token).

On the client side, use -u <username> to set the login and --password to be prompted for the server’s password. It is also possible to set the client passwords in the Glances configuration file:

[passwords]
# Define the passwords list
# Syntax: host=password
# Where: host is the hostname
#        password is the clear password
# Additionally (and optionally) a default password could be defined
localhost=mylocalhostpassword
default=mydefaultpassword

Fetch mode

It is also possible to get and share a quick look of a machine using the fetch mode. In this mode, current stats are display on the console in a fancy way.

$ glances --fetch

Results look like this:

_images/screenshot-fetch.png

It is also possible to use a custom template with the --fetch --fetch-template </path/to/template.jinja> options.

Have a look to the fetch documentation page to learn how to create your own template.