The most common ways to deploy Efesto are by using a WSGI server of your choice (suchs as gunicorn or uwsgi) or by using the available docker images.

If you are using a WSGI server, all you need to do is to set the options in the enviroment.

Docker-based deploy

Efesto can be found as docker image under strangemachines/efesto. The images use gunicorn as WSGI server.

To facilitate docker-based deployments and allow passing options to gunicorn, there are a number of additional environment options available.


When given, a superuser with the given name will be created.


The BLUEPRINT environment variable is a way to pass to the efesto container the blueprint that you want to use, instead of having to load that manually.

It's common to set this value with cat: export BLUEPRINT=$(cat <your-blueprint>)


The port on which gunicorn will run. Defaults to :5000


How many workers should gunicorn use. Defaults to 3.


The class of the gunicorn workers. Defaults to sync. Can be sync, tornado, gevent or meinheld. Meinheld requires the specific build.

When set, it will also install the worker with pip install.


How many threads should gunicorn use. Defaults to 1.


The timeout before a worker is restarted. It's passed to both -t and --graceful-timeout. Defaults to 60.


Gunicorn jitter to randomize workers timeout time. Defaults to $WORKERS * 5.

results matching ""

    No results matching ""