Server deployment

To deploy the service, you can either use our provided Docker images (recommended) or setup the dependencies manually. The server requires a MongoDB instance, that is not included in the Docker container.

Docker

We provide a docker image with the API server. You can either pull a prebuilt image from the kobzol/snailwatch repository on DockerHub or build it yourself using the Dockerfile located in the server directory.

# pull prebuilt image
$ docker pull kobzol/snailwatch:server

# or build it yourself
$ docker build -t sw-server -f server/Dockerfile

The app inside the container is located at /server and is by default server on port 5000.

Manual installation

You can install the dependencies of the server with pip. Snailwatch should be compatible with Python 2 and 3:

$ cd server
$ pip install -r requirements.txt

Configuration

The server can be configured with the following environment variables:

Name	Required	Default	Description
ADMIN_TOKEN	True		Authentication token used for administrative actions.
MONGO_HOST	False	localhost	Address of MongoDB server.
MONGO_PORT	False	27017	Port of MongoDB server.
MONGO_DB	False	snailwatch	Name of MongoDB database.
MONGO_USERNAME	False		Auth username for MongoDB server.
MONGO_PASSWORD	False		Auth password for MongoDB server.
SMTP_SERVER	False	localhost	SMTP server for sending regression notifications.
SMTP_USER	False		SMTP user.
SMTP_PASSWORD	False		SMTP password.
EMAIL_SENDER	False	{SMTP_USER}	Appears as a sender of notification e-mails.
PUBLIC_URL	False		Public URL where Snailwatch is hosted (used for links in notification e-mails).
PORT	False	5000	TCP port used by the API server.

Running the server

After you install and configure the server, you have to start the MongoDB instance and afterwards you can launch the server itself using:

# Start using Docker (MongoDB instance has to be available in its network)
$ docker run -e "ADMIN_TOKEN=abc" -p 5000:5000 sw-server

# Start directly
$ python start.py