Skip to content

Docker Compose Advanced Use Cases

This section describes some of the more advanced use cases for the Docker Compose development workflow.

Invoke Configuration

The Invoke tasks have some default configuration which you may want to override. Configuration properties include:

  • project_name: The name that all Docker containers will be grouped together under (default: nautobot, resulting in containers named nautobot_nautobot_1, nautobot_redis_1, etc.)
  • python_ver: the Python version which is used to build the Docker container (default: 3.8)
  • local: run the commands in the local environment vs the Docker container (default: False)
  • compose_dir: the full path to the directory containing the Docker Compose YAML files (default: "<nautobot source directory>/development")
  • compose_files: the Docker Compose YAML file(s) to use (default: ["docker-compose.yml", "docker-compose.postgres.yml", ""])
  • docker_image_names_main and docker_image_names_develop: Used when building Docker images for publication; you shouldn't generally need to change these.

These setting may be overridden several different ways (from highest to lowest precedence):

  • Command line argument on the individual commands (see invoke $command --help) if available
  • Using environment variables such as INVOKE_NAUTOBOT_PYTHON_VER; the variables are prefixed with INVOKE_NAUTOBOT_ and must be uppercase; note that Invoke does not presently support environment variable overriding of list properties such as compose_files.
  • Using an invoke.yml file (see invoke.yml.example)

Working with Docker Compose

The files related to the Docker development environment can be found inside of the development directory at the root of the project.

In this directory you'll find the following core files:

  • docker-compose.yml - Docker service containers and their relationships to the Nautobot container
  • docker-compose.debug.yml - Docker compose override file used to start the Nautobot container for use with Visual Studio Code's dev container integration.
  • - Docker compose override file used to mount the Nautobot source code inside the container at /source and the from the same directory as /opt/nautobot/ for the active configuration.
  • - Docker compose override file used to start/build the final (production) Docker images for local testing.
  • - Docker compose override file used to start/build the final-dev (app development environment) Docker images for local testing.
  • docker-compose.keycloak.yml - Docker compose override file used to setup an SSO auth backend for Nautobot.
  • docker-compose.mysql.yml - Docker compose override file used to add a MySQL container as the database backend for Nautobot.
  • docker-compose.postgres.yml - Docker compose override file used to add a Postgres container as the database backend for Nautobot.
  • dev.env - Environment variables used to setup the container services
  • - Nautobot configuration file

In addition to the development directory, additional non-development-specific Docker-related files are located in the docker directory at the root of the project.

In the docker directory you will find the following files:

  • Dockerfile - Docker container definition for Nautobot containers
  • - Commands and operations ran once Nautobot container is started including database migrations and optionally creating a superuser
  • uwsgi.ini - The uWSGI ini file used in the production docker container

Docker Compose Overrides

If you require changing any of the defaults found in docker-compose.yml, create a file inside the development directory called docker-compose.override.yml and add this file to the compose_files setting in your invoke.yml file, for example:

    - "docker-compose.yml"
    - "docker-compose.postgres.yml"
    - ""
    - "docker-compose.override.yml"

This file will override any configuration in the main docker-compose.yml file, without making changes to the repository.

Please see the official documentation on extending Docker Compose for more information.

Automatically Creating a Superuser

There may be times where you want to bootstrap Nautobot with a superuser account and API token already created for quick access or for running within a CI/CD pipeline. By using a custom invoke.yml as described above, in combination with custom docker-compose.override.yml and override.env files, you can automatically bootstrap Nautobot with a user and token.

Create invoke.yml as described above, then create development/docker-compose.override.yml with the following contents:

      - "override.env"

The script will run any migrations and then look for specific variables set to create the superuser. The script is copied in during the Docker image build and will read from the default dev.env as the env_file until you override it as seen above.

Any variables defined in this file will override the defaults. The override.env should be located in the development/ directory, and should look like the following:

# Superuser information. NAUTOBOT_CREATE_SUPERUSER defaults to false.

The variables defined above within override.env will signal the script to create the superuser with the specified username, email, password, and API token.

After these two files are created, you can use the invoke tasks to manage the development containers.

Using MySQL instead of PostgreSQL

By default the Docker development environment is configured to use a PostgreSQL container as the database backend. For development or testing purposes, you might optionally choose to use MySQL instead. In order to do so, you need to make the following changes to your environment:

  • Set up invoke.yml as described above and use it to override the postgres docker compose file:
    - "docker-compose.yml"
    - "docker-compose.mysql.yml"
    - ""

Then invoke stop (if you previously had the docker environment running with Postgres) and invoke start and you should now be running with MySQL.

SSO Auth Backend with Keycloak

Keycloak and its database are run in the same docker compose project as Nautobot. A separate database is used to ensure you are able to have two separate instances of Postgres, one for Nautobot and one for Keycloak, and allows you to use a MySQL database for Nautobot but maintain Keycloaks required Postgres DB. This setup is meant for local development and testing, and should not be used as a reference for deploying Keycloak in production.

The invoke.yml file must be updated to add development/docker-compose.keycloak.yml to the docker compose project and to enable OIDC. These setting are solely for local development inside the Nautobot repository and is not applicable to any other deployment. An example invoke.yml file:

    - "docker-compose.yml"
    - "docker-compose.postgres.yml"
    - ""
    - "docker-compose.keycloak.yml"

Validating Setup

Once all steps are completed Nautobot should now have the Continue to SSO button on the login screen and should immediately redirect the user to sign in with Keycloak.

Keycloak Login Credentials

Keycloak admin console is reachable via http://localhost:8087/admin/ with user admin and password admin. The below users are pre-configured within Keycloak, at this time their permissions are not directly mapped to any permissions provided by default by Nautobot. This will be a later enhancement to the local development environment.

| Username | Password | +------------------+-----------+ | nautobot_unpriv | unpriv123 | | nautobot_admin | admin123 | | nautobot_auditor | audit123 |

Microsoft Visual Studio Code Integration

For users of Microsoft Visual Studio Code, several files are included to ease development and integrate with the VS Code Remote - Containers extension. The following related files are found relative to the root of the project:

  • .devcontainers/devcontainer.json - Dev. container definition
  • nautobot.code-workspace - VS Code workspace configuration for Nautobot
  • development/docker-compose.debug.yml - Docker Compose file with debug configuration for VS Code

After opening the project directory in VS Code in a supported environment, you will be prompted by VS Code to Reopen in Container and Open Workspace. Select Reopen in Container to build and start the development containers. Once your window is connected to the container, you can open the workspace which enables support for Run/Debug.

To start Nautobot, select Run Without Debugging or Start Debugging from the Run menu. Once Nautobot has started, you will be prompted to open a browser to connect to Nautobot.


You can run tests with nautobot-server --config=nautobot/core/tests/ test nautobot while inside the Container.

Special Workflow for Containers on Remote Servers

A slightly different workflow is needed when your development container is running on a remotely-connected server (such as with SSH). VS Code will not offer the Reopen in Container option on a remote server.

To work with remote containers, after invoke build use docker-compose as follows to start the containers. This prevents the HTTP service from automatically starting inside the container:

cd development
docker compose -f docker-compose.yml -f docker-compose.debug.yml up
  • Now open the VS Code Docker extension. In the CONTAINERS/development section, right click on a running container and select the Attach Visual Studio Code menu item.
  • The Select the container to attach VS Code input field provides a list of running containers.
  • Click on development_nautobot_1 to use VS Code inside the container. The devcontainer will startup now.
  • As a last step open the folder /opt/nautobot in VS Code.

Remote Debugging Configuration

Using the Remote-Attach functionality of VS Code debugger is an alternative to debugging in a development container. This allows a local VS Code instance to connect to a remote container and debug the code running in the container the same way as when debugging locally.

Follow the steps below to configure VS Code to debug Nautobot and Celery Worker running in a remote container:

  1. Configure invoke.yml to use the docker-compose.vscode-rdb.yml file

    This overrides the container settings without touching the original docker-compose.yml file.

    Your invoke.yml file should look something like this:

        - "docker-compose.yml"
        - "docker-compose.postgres.yml"
        - ""
        - "docker-compose.vscode-rdb.yml"

    If you already have a custom invoke.yml file, append docker-compose.vscode-rdb.yml to the end of the compose_files list.

    See the docker compose override documentation for more details.

  2. VS Code debug configuration

    If you have opened the workspace file nautobot.code-workspace then there are two debug configurations for remote debugging already available.

    If you add nautobot to an existing VS Code workspace (Menu: File > Add Folder to Workspace...) then copy the launch: values to the .vscode/launch.json file.

    • Debug Configurations for Remote Debugging:
      "version": "0.2.0",
      "configurations": [
          "name": "Python: Remote Attach",
          "name": "Python: Remote Attach Celery Worker",

It is now possible to debug the containerized Nautobot and Celery Worker using the VS Code debugger.

After restarting the Celery-Worker container you need to restart the debug session.