Manager Configuration¶

The Manager is highly configurable using environment variables.

Environment Variable	Default Value	Description
MVI_HOST_NAME	“localhost”	Host name of the machine running the Manager.
MVI_PROXY_HTTP_PORT	80	Proxy port for HTTP communication with Manager.
MVI_PROXY_HTTPS_PORT	443	Proxy port for HTTP communication with Manager.
MVI_PROXY_CONFIG_PATH	“proxy_config”	Path to proxy configuration files.
MVI_AUTH_ENABLED	False	Using the Manager requires login if true.
MVI_LOG_LEVEL	“INFO”	Minimum logger level for Manager.
MVI_ACCESS_LOGS_ENABLED	True	Whether to print access logs.
MVI_PROXY_HTTPS	False	Sets up a secure connection to the Manager if true
MVI_HTTPS_STAGING_SERVER	False	Sets if a staging server without rate limits should be used
MVI_CONFIG_EMAIL	null	Email address HTTPS and sending notifications.
MVI_CONFIG_EMAIL_PASSWORD	null	Password for config email, only required for notifications.
MVI_NOTIFICATION_SMTP_SERVER	null	SMTP server URL.
MVI_NOTIFICATION_SMTP_PORT	null	Port to SMTP server, usually 587 or 465.
MVI_ACTIVATION_KEY	“”	License activation key. Without key Manager lives for 12h.

Secure Manager Connection¶

It is possible to get a secure HTTPS connection to the Manager using automatically created certificates from Let’s Encrypt. To enable HTTPS you must start a new Manager with the environment variable MVI_PROXY_HTTPS set to true. We also recommend that you set an email address with the environment variable MVI_CONFIG_EMAIL, so you can get notified if the certificate is about to run out.

Warning

For the Let’s Encrypt certificates to work, the server must have a valid hostname which resolves via DNS to the server’s IP address. It will not work if you access the Manager using an IP address or localhost due to how TLS certificates are generated.

Email notifications¶

The first step is to set up an email address that the MVI Manager can use to send the notification emails. For this you need an SMTP server with an email account that you can use. We recommend to have a dedicated email address for sending MVI notifications.

The Manager is configured with a notification sender email at startup. We do this by setting four different environment variables at Manager startup

Environment Variable	Value
MVI_CONFIG_EMAIL	“email@address.com”
MVI_CONFIG_EMAIL_PASSWORD	“password”
MVI_NOTIFICATION_SMTP_SERVER	“smtp.server.url”
MVI_NOTIFICATION_SMTP_PORT	Usually 587 or 465

Upon startup of the Manager, if everything worked, an email will be sent from “email@address.com” to itself, to show that it successfully connected. Otherwise, as long as the Manager could still start, any issues with the email configuration will be viewable on the notification tab in the dashboard.

Log configuration¶

All logging in MVI is done directly to stdout and stderr and relies heavily on the built-in logging features of the docker daemon. As such, any configuration of log rotation etc needs to be done on the docker daemon level.

MVI takes care of setting reasonable log configuration options on all the services that are started. For each service, this means that log files are rotated when they grow above 100MB in size and a maximum of 3 such files are kept on disk.

For the Manager container however, the log configuration is left to the user. But we give some hints here on reasonable options. In general, the docker daemon and its logging mechanism can be configured in two ways, either by editing the daemon.json configuration file (which sets daemon-wide default configurations) or by providing container-specific configuration when starting a new container with docker run. For more details, please have a look at the docker docs.

An example container-specific configuration could be provided as such (maximum of 5 log files no larger than 10 megabytes each using the JSON log handler (the docker daemon default)):

docker run ... --log-driver json-file --log-opt max-size=100m --log-opt max-file=5

Warning

By default, the docker daemon is configured to NOT do any log rotation at all, meaning that it will slowly fill up the HDD of the host. To avoid any problems originating from a full HDD, we highly recommend setting a specific log configuration for the Manager container when starting.

Typical production setup¶

The example below starts a Manager instance that is listening on my.domain.com, enforcing traffic over https with access control enabled and being activated using a provided activation key:

# We create a docker volume for keeping our data persistent across restarts/upgrades
docker volume create mvi_data

docker run \
    --name mvi_manager \
    -v /var/run/docker.sock:/var/run/docker.sock \
    -v mvi_data:/data \
    -p 80:80 \
    -p 443:443 \
    -e MVI_HOST_NAME=my.domain.com \
    -e MVI_PROXY_HTTPS=True \
    -e MVI_AUTH_ENABLED=True \
    -e MVI_ACTIVATION_KEY=... \
    -e MVI_CONFIG_EMAIL=<some@email.com> \
    -e MVI_CONFIG_EMAIL_PASSWORD=<password for some@email.com> \
    -e MVI_NOTIFICATION_SMTP_SERVER=<your smtp server address> \
    -e MVI_NOTIFICATION_SMTP_PORT=<your smtp server port> \
    --restart always \
    --log-driver json-file \
    --log-opt max-size=100m \
    --log-opt max-file=5 \
    -d ghcr.io/vikinganalytics/mvi/mvi_manager:{version} \

Tip

To ease working with all environment variables, it is possible to make use of the --env-file parameter to docker run. See here for details on syntax etc.