Docker

Let’s create an installation directory under a docker folder in our servers home folder

mkdir -p ~/docker/monitee-agent/{config,data} && cd ~/docker/monitee-agent

The server expects a configuration.yml in the configuration directory. Let’s download the latest one using wget

cd config

wget https://raw.githubusercontent.com/Krillsson/monitee-agent/refs/heads/master/config/configuration.yml
wget https://raw.githubusercontent.com/Krillsson/monitee-agent/refs/heads/master/config/application.properties

Use your favorite text editor to edit the user & password of the config

nano configuration.yml

Then press CTRL+X to exit nano (and press Y to save)

Now we can start the container.

docker run -d \
    --net=host \
    --pid=host \
    --name monitee-agent \
    --cap-add SYS_RAWIO \
    -v ~/docker/monitee-agent/data:/data \
    -v ~/docker/monitee-agent/config:/config \
    -v /etc/localtime:/etc/localtime:ro \
    -v /var/run/docker.sock:/var/run/docker.sock \
    -v /run/udev:/run/udev:ro \
    -v /dev:/dev:ro \
    -v /srv:/srv:ro \
    -v /etc/os-release:/etc/os-release:ro \
    krillsson/sys-api

Want to monitor the temperature of disks? Each device you’d like to monitor has to be added to the command above through –device=/dev/sdX (sda, sdb, etc). List disks on your system with lsblk –nodeps -n -o name

Want to monitor a nvidia GPU? Also add this to the command:
-e ‘NVIDIA_VISIBLE_DEVICES’=’all’
-e ‘NVIDIA_DRIVER_CAPABILITIES’=’utility’
–runtime=nvidia

or substitute all with the specific gpu index (e.g “0”)

The agent runs the native build, which uses less memory, and falls back to Java on hosts it can’t run on. Want to choose for yourself? Add -e SYSAPI_RUNTIME=jvm to always run Java, or -e SYSAPI_RUNTIME=native to run the native build without falling back.

To manage systemd services and read the journal from the app, the host has to run systemd and the container needs these mounts as well. They are left out above because not every Docker host uses systemd. Add them to the command:

    -v /run/systemd:/run/systemd \
    -v /etc/machine-id:/etc/machine-id:ro \
    -v /run/systemd/journal/socket:/run/systemd/journal/socket:ro \
    -v /run/log/journal:/run/log/journal:ro \

Hosts that don’t use systemd can skip this. The agent turns the features off on its own when systemd isn’t reachable. They can also be switched off under linux in configuration.yml.

The file browser is off by default. It can only reach the directories listed under roots in configuration.yml, so a container needs the host directory bind mounted first. Add the mount to the command above:

    -v /mnt/user:/storage \

and then list the path inside the container in configuration.yml:

fileBrowser:
  enabled: true
  access: READ
  roots:
    - /storage

access: READ gives browsing, reading and downloading. access: READ_WRITE also allows editing, uploading, copying, moving and deleting. A few things worth knowing before turning that on:

• The container runs as root, so READ_WRITE means the agent password is write access to everything you mount. Mount the directories you actually want to reach, never /.
• Mount the volume :ro while you are on access: READ, and the kernel enforces the read only part for you.
• Keep /config and /data out of the roots. If they are reachable, READ_WRITE lets a caller edit the agent’s own configuration and database, and the agent says so in the log at startup.

Let’s verify the server started correctly

docker logs --follow --tail 200 monitee-agent

When you see the following, we know that the server is up and running:

[...]: Tomcat started on ports 8443 (https), 8080 (http) with context path '/'
[...]: Started SysAPIApplicationKt in 6.352 seconds (process running for 7.234)

Now you are ready to proceed to next step

Connect App to Server