Skip to main content

Glances smart home integration, free and open source

Glances integration for Gladys Assistant

Follow your NAS, servers and PCs with Glances: CPU, memory, disks, temperatures, containers.

View on GitHubAll external integrations

Glances is a cross-platform monitoring tool: CPU, memory, disks, network, temperatures, Docker and Podman containers. This integration reads the Glances web server of your other machines — a NAS, a home server, a PC — and brings them into Gladys: one device per server, and the disks, network interfaces, temperature sensors and containers you choose. Scenes can react when a disk fills up, a machine overheats, a container stops or a server no longer answers, and a dashboard widget sums a server up. Everything stays on your local network.

Tested against real Glances 4.5.8 and 3.4.0.5 servers (Docker), with and without a password; the containers through a simulated Docker engine. Not yet on every NAS: feedback welcome on the issues.

What you need​

  • Gladys Assistant 5.1.0 or later.
  • On each machine to follow, Glances 3.x or 4.x started as a web server (glances -w), reachable from the Gladys machine (port 61208 by default).
  • Up to three machines per installation of the integration.

The integration only reads: it never changes anything on your machines.

Getting started​

1. Start Glances in web server mode​

On the machine to follow:

  • installed with pip or your distribution: glances -w (as a service so it survives a reboot); add --password to protect it, Glances asks for the password once and can save it;
  • with Docker: the nicolargo/glances image with GLANCES_OPT="-w", port 61208 published, and, to see the containers, the Docker socket mounted read-only (-v /var/run/docker.sock:/var/run/docker.sock:ro). To see the disks of the host, mount its root read-only (-v /:/rootfs:ro).

Check from a browser that http://<machine address>:61208 shows Glances.

2. Configure the integration​

In Integrations → Glances → Configuration:

  • Glances address: 192.168.1.20, nas.lan, or the address copied from your browser (http://nas.lan:61208); port 61208 is added when none is given;
  • User name (optional) and Password (optional): only if Glances was started with --password (the user name is glances unless Glances was started with --username);
  • Name (optional): the name shown in the devices, the widget and the messages; by default, the host name Glances gives (or the address when Glances runs in Docker and only knows its container id);
  • the same for a second and a third server, if you have them;
  • Refresh interval: 15 seconds to 5 minutes (30 seconds by default);
  • Accept self-signed certificates: only for an https:// address (Glances behind a reverse proxy) whose certificate is self-signed;
  • Language of the names: the language of the feature names (set when a device is added), of the scene results and of the connection test (French by default).

Test the connection reads every server now and says what it found: the Glances version, the number of disks, interfaces, sensors and running containers.

3. Add the devices​

Open the Discover tab: the integration proposes the server itself, then each of its disks, network interfaces, temperature sensors and containers. Add only what you want to follow: a NAS easily has thirty of them. The scene triggers work for every disk, sensor and container, added or not.

The devices​

Values are rounded and sent only when they change; the ones that move all the time (CPU, load, rates, temperatures, containers) at most once a minute.

The server​

FeatureValue
OnlineOn while Glances answers, Off after two silent reads
CPUUse of all the cores, in %
MemoryMemory used, in %
SwapSwap used, in % (when the machine has swap)
Load (1 min)Load average over 1 minute
Load (5 min)Load average over 5 minutes
Load (15 min)Load average over 15 minutes
UptimeTime since the last boot, in days
Running containersDocker or Podman containers running (when there are any)
Highest temperatureThe hottest temperature sensor, in °C

Disks​

One device per mount point: Space used (%) and Free space (GB). The files Docker binds into a Glances container and the system pseudo file systems are left out. Glances 4 measures the disks once a minute.

Network interfaces​

One device per interface: Download and Upload, in Mbit/s. Loopback, Docker bridges and virtual interfaces (veth…, br-…, docker0…) are left out.

Temperature sensors​

One device per temperature sensor Glances reads (Core 0, Package id 0, Composite for an NVMe drive…): Temperature, in °C.

Containers​

One device per container, kept when the container is recreated with the same name (a docker compose update): Running (On/Off), CPU (%, one core counts 100 %, like docker stats) and Memory (MB, page cache left out). By default Glances only lists the running containers: a stopped container disappears from Glances, and its device goes Off after two reads without it. To list the stopped ones too, set all=True in the [containers] section of glances.conf.

Dashboard​

Add the Glances server widget to a dashboard and choose the server in its settings (any device of the server will do: the server itself, one of its disks…). It shows:

  • three gauges: CPU, memory and the disk (the root disk, or the disk device chosen in the settings), orange from 75 %, red from 90 %;
  • the highest temperature (orange from 70 °C, red from 85 °C) and the number of running containers (red when one stopped);
  • rows: the containers that stopped (in red, first), the free space of the disk, the swap, the load and the uptime;
  • a server that does not answer, and why;
  • a Refresh button that reads the server again.

Scenes​

Triggers​

They fire for every disk, sensor and container, added to Gladys or not. The Server field narrows a trigger to one server (choose the device of the server itself); leave it empty for all.

  • Disk almost full: a disk reaches the chosen use (80, 85, 90, 95 or 98 %). Once per step; again only after the disk went 2 points below it. Optional filter: one disk. Variables: server, mount point, space used (%), free space (GB), threshold.
  • High temperature: a sensor stays at or above the chosen temperature (50 to 100 °C) for two reads in a row. Again only after it went 5 °C below, and not twice within 30 minutes. Optional filter: one sensor. Variables: server, sensor, temperature, threshold.
  • Container stopped: a container that was running stops, is paused or keeps restarting, on two reads in a row (a container recreated by an update is no stop). Optional filter: one container. Variables: server, container, state ("Stopped", "Paused", "Restarting").
  • Server unreachable: a server no longer answers (two reads in a row, 15 seconds apart). A refused password is not announced: the connection status shows it. Variables: server, reason.

Each threshold is a filter: one scene per threshold you care about. What was announced is kept, so a restart of Gladys or of the integration neither repeats an announcement nor misses one. A disk or a sensor already above a threshold when it is seen for the first time announces nothing.

Actions​

  • State of a server: reads a server for the next actions of the scene: name, answers (yes/no), CPU, memory and swap (%), load averages, uptime (days), main disk used (%) and free (GB), highest temperature, running and stopped containers, and a ready-made summary such as "NAS: CPU 12%, memory 45%, disk / 78% (110 GB free), 52 °C, 12 container(s) running". A server that does not answer is not an error: "answers" is no and the summary says why.

It reads the server again when the last read is older than 15 seconds.

Scene ideas​

  • Disk alert: trigger "Disk almost full" with the threshold 90 % → send a message "💾 The disk … of … is at …%", inserting the variables "Mount point", "Server" and "Space used (%)".
  • Morning report: every day at 8:00 → "State of a server" on the NAS → send the "Summary" output by message.
  • A service fell: trigger "Container stopped" on the container of Nextcloud → send a message, or turn a light red.
  • Overheat: trigger "High temperature" at 85 °C → send a message, switch on the fan of the server cupboard (a smart plug).

Troubleshooting​

  • "does not answer": Glances is not started in web server mode, the port is not 61208, or a firewall blocks it. Open http://<address>:61208 from another machine.
  • "user name or password refused": check the password of glances -w --password, and the user name if Glances was started with --username.
  • "Glances asks for a password": Glances was started with --password; enter it in the configuration.
  • "no Glances web server (API 3 or 4) at this address": another service answers on that port, or Glances is older than 3.0.
  • No container in the Discover tab: Glances does not see the container engine. In Docker, mount the Docker socket into the Glances container.
  • A disk is missing: Glances hides /boot and /snap, and does not list tmpfs and network shares by default (see the [fs] section of glances.conf). In Docker, mount the host root (-v /:/rootfs:ro).
  • The CPU does not move for a while: Glances 4 sometimes answers a CPU of zero everywhere (at its first read, or after a pause); that reading is skipped, the value comes at the next one.

Limits​

  • Read only: the integration starts or stops nothing, neither containers nor machines.
  • Three servers at most per installation.
  • The values of the devices that are not added are not sent to Gladys; the triggers and the widget still use them.
  • A container removed for good is forgotten after a week.
  • GPU, RAID, disk I/O, processes, fans and batteries are not read.
  • Not affiliated with the Glances project.

Configuration settings​

These are the settings Glances asks for in its configuration screen in Gladys.

SettingTypeRequiredDescription
Your Glances serverssectionNoOn each machine to follow (NAS, server, PC), run Glances in web server mode: "glances -w", or the nicolargo/glances Docker image with GLANCES_OPT="-w". Enter its address below, then add the server, and the disks, interfaces, sensors and containers you want, from the Discover tab.
Glances addressstringNoAddress of the machine that runs "glances -w" (port 61208 by default): 192.168.1.20, nas.lan or http://nas.lan:61208.
User name (optional)stringNoOnly with a password: "glances", unless Glances was started with --username.
Password (optional)secretNoThe password of "glances -w --password". Leave empty when Glances asks for none.
Name (optional)stringNoShown in the device names, the widget and the messages. By default, the host name Glances gives.
Second Glances server (optional)sectionNoAnother machine, with its own address and password.
Second Glances addressstringNo
Second user name (optional)stringNo
Second password (optional)secretNo
Second name (optional)stringNo
Third Glances server (optional)sectionNoAnother machine, with its own address and password.
Third Glances addressstringNo
Third user name (optional)stringNo
Third password (optional)secretNo
Third name (optional)stringNo
OptionssectionNoOnly the devices added to Gladys receive values; the scene triggers work for every disk, sensor and container.
Refresh intervalselectNoHow often every Glances server is read.
Accept self-signed certificatesbooleanNoOnly for an https:// address (Glances behind a reverse proxy) whose certificate your browser does not trust either.
Language of the namesselectNoLanguage of the feature names (set when a device is added), of the scene results and of the connection test.

How to install Glances in Gladys​

  1. In Gladys, open Integrations: Glances appears in the catalog, next to the native integrations, with a community badge.
  2. Click Install. Gladys pulls the Docker image (ghcr.io/guim31/gladys-glances:1.0.1), starts it in a sandbox isolated from the core, and generates the integration's interface (devices, discovery and configuration).
  3. Open the Configuration screen of the integration, fill in the settings, and save.
  4. You can also install it directly from its repository URL: https://github.com/guim31/gladys-glances.

Glances requires Gladys >=5.1.0. The catalog inside Gladys refreshes every hour, so a new version becomes available at most one hour after its release.

Not running Gladys yet? It is free and open source: follow the installation guide to get started.

About external integrations​

Glances is an external integration: a community integration packaged as a Docker container and published on GitHub, that Gladys installs in one click and runs in a sandbox isolated from its core. It is published and maintained by guim31, not by the Gladys core team.

Subscribe to the Gladys Assistant newsletter

A few emails per month about new releases and project news. Sent by Pierre-Gilles Leymarie, founder of the project. Unsubscribe anytime 🙂
Last updated on