Install Mideye Fleet with Docker: Quickstart or Image

Docker on Ubuntu is the supported way to run Mideye Fleet in this version; Fleet on Windows Server is planned for the next version. Pick one of two paths.

Path	Use it when
Docker quickstart	You want Mideye Server and Fleet together on one Ubuntu host. This is the fastest start and the best choice for a first test.
Fleet image on its own	You already run Mideye Server and want to add Fleet.

Before you start, read the requirements.

Path 1: Docker quickstart

The Docker quickstart puts four containers on one Ubuntu host: MariaDB, Mideye Server, a one-time setup container and Mideye Fleet. You run one command, fill in one file and start it. The setup container prepares Mideye Server for Fleet, so Fleet already knows your Mideye Server and the endpoint installer when you first sign in.

In short:

Sign in to the download portal and open the Fleet tab, Docker section. Run the docker login line you find there.
Run the install command from the quickstart. It prints the credentials for the first sign-in. Store them in your password manager. They are shown once.
Fill in /opt/mideye/.env: the host name, the Mideye Server password and, if you use Mideye Cloud, the cloud credentials.
Start the containers with docker compose up -d.
Open https://<your host name>:9443 and continue with the first sign-in, as the quickstart describes.

If root loses its password or authenticator later, see Reset root or an admin.

The quickstart page has every command and all fields. This page does not repeat them. Then add your first Windows computer with the Windows quickstart, and let users set up an authenticator app with the Windows TOTP quickstart.

Path 2: The Fleet image on its own

Use this when Mideye Server already exists. Download the Fleet Docker image from the download portal and check it against the checksum on the portal.

Create a folder for the configuration and a Docker volume for the data:
Terminal window
```
docker volume create fleet-data
mkdir -p /srv/mideye-fleet/cfg
```
Write /srv/mideye-fleet/cfg/fleet.yaml. Put in the name that administrators and computers use to reach Fleet. Computers check this name against Fleet’s certificate.
```
data_dir: /data
listen: ":9443"
tls:
  hostnames:
    - fleet.example.com
database:
  driver: sqlite
```
SQLite needs nothing else. For MariaDB, MySQL or SQL Server, see the requirements and the Fleet guides in the Mideye downloads portal (login required).

Start the container. Always give /data a volume.

docker run -d --name mideye-fleet \
  --read-only \
  --tmpfs /tmp:rw,noexec,nosuid,size=64m \
  --cap-drop ALL \
  --security-opt no-new-privileges \
  --restart unless-stopped \
  -v fleet-data:/data \
  -v /srv/mideye-fleet/cfg:/cfg:ro \
  -p 9443:9443 -p 192.0.2.10:9444:9444 \
  mideye-fleet-server:<version> --config /cfg/fleet.yaml

Replace 192.0.2.10 with the address that your computers use to reach Fleet. Docker publishes ports on every interface and goes around the host firewall, so give port 9444 an address.

Open https://<host>:9443 and continue with the first-run steps in the Fleet guides in the Mideye downloads portal (login required).

If the container stops

Run docker logs mideye-fleet --tail 50. Exit code 78 means Fleet needs you to fix something first, for example a wrong setting in fleet.yaml or a data folder with the wrong permissions. Fleet says what to fix. Docker with --restart unless-stopped starts it again and the same message repeats until you fix the cause.

Before you add computers

Open Settings and set the Fleet Server URLs. These are the addresses that computers use to reach Fleet on port 9444, for example https://fleet.example.com:9444. The quickstart sets it for you.
Use a DNS name, not an IP address. Moving Fleet later is much easier when computers know it by name.

The Fleet installation and operation guides are in the Mideye downloads portal (login required).

Other packages

The download portal also lists Fleet packages for Windows Server (MSI) and for Debian, Ubuntu, RHEL and Rocky Linux (deb and rpm), marked as preview. Install guides for them are coming. Until then, use Docker.