Running via Docker
FlexMeasures can be run via docker.
Docker is great to save developers from installation trouble, but also for running FlexMeasures inside modern cloud environments in a scalable manner.
We also support running all needed parts of a FlexMeasures EMS setup via docker-compose, which is helpful for developers and might inform hosting efforts. See Running a complete stack with docker-compose.
For now, the use case is local development. Using in production is a goal for later. Follow our progress.
The flexmeasures image
Getting the image
You can use versions we host at Docker Hub, e.g.:
docker pull lfenergy/flexmeasures:latest
You can also build the FlexMeasures image yourself, from source:
docker build -t flexmeasures/my-version .
The tag is your choice.
Running the image (as a container) might work like this (remember to get the image first, see above):
docker run --env SQLALCHEMY_DATABASE_URI=postgresql://user:pass@localhost:5432/dbname --env SECRET_KEY=blabla --env FLASK_ENV=development -d --net=host lfenergy/flexmeasures
Don’t know what your image is called (its “tag”)? We used
lfenergy/flexmeasures here, as that should be the name when pulling it from Docker Hub. You can run
docker images to see which images you have.
The two minimal environment variables to run the container successfully are the database URI and the secret key, see Configuration.
FLASK_ENV=development is needed if you do not have an SSL certificate set up (the default mode is
production, and in that mode FlexMeasures requires https for security reasons). If you see too much output, you can also set
In this example, we connect to a postgres database running on our local computer, so we use the host network. In the docker-compose section below, we use a Docker container for the database, as well.
http://localhost:5000 should work now and ask you to log in.
Of course, you might not have created a user. You can use
docker exec -it <flexmeasures-container-name> bash to go inside the container and use the CLI Commands to create everything you need.
Configuration and customization
Using Configuration by file is usually what you want to do. It’s easier than adding environment variables to
docker run. Also, not all settings can be given via environment variables. A good example is the MAPBOX_ACCESS_TOKEN, so you can load maps on the dashboard.
To load a configuration file into the container when starting up, we make use of the instance folder. You can put a configuration file called
flexmeasures.cfg into a local folder called
flexmeasures-instance and then mount that folder into the container, like this:
docker run -v $(pwd)/flexmeasures-instance:/app/instance:ro -d --net=host lfenergy/flexmeasures
The location of the instance folder depends on how we serve FlexMeasures. The above works with gunicorn. See the compose file for an alternative (for the FlexMeasures CLI), and you can also read the above link about the instance folder.
This is also a way to add your custom logic (as described in Writing Plugins) to the container. We’ll document that shortly. Plugins which should be installed (e.g. by
pip) are a bit more difficult to support (you’d need to add pip install before the actual entry point). Ideas welcome.