Understanding Kargo

WARNING

Kargo mainly relies on Docker Swarm (opens new window). We assume that you are enough familiar with this technology. If not, you should take a little time to learn about it !

Key concepts

Kargo let you manipulate 4 kind of entities:

• configuration: a collection of everything Kargo needs to create and manage the services to deploy on the infrastructure.
• Environment: a set of environment variables used to specify the stacks of services to deploy and how to deploy them.
• Stack: a group of interrelated services that can be orchestrated and scaled together.
• Service: an application or a piece of application deployed on your Cluster.

The following image illustrates how these entities interact:

Configuration

As presented in the previous diagram, we distinguish three types of configuration:

• the default configuration: by default Kargo provides a mean to deploy preconfigure services. The settings for these services are often minimalist and you may need to overwrite them to configure the services as you wish.
• the user configuration: a directory where you store the Environment file as well as the settings of the different services you want to overwrite.
• the merged configuration: a directory where Kargo merges the two previous configuration. The directory is named .kargo and is located at the root of the Kargo folder.

TIP

The user configuration may store some sensitive data (passwords, api access keys...) thus we recommend to store your it on a private repository.

WARNING

The merged configuration must be generated before building or deploying the services. See the subcommand configure.

Environment

The environment is group of environment variables declared in a file named .env. It allows to define:

• the main settings of Kargo
• the stacks of services to deploy
• the deployment settings for each service

Stack

A stack of interrelated services that can be orchestrated and scaled together. Kargo provides some commands to deploy and remove the stacks independently.

Service

A service is an application to be deployed over the Swarm cluster. Kargo provides the default settings for deploying the following services/applications:

• Adminer (opens new window)
• Centipede RTK (opens new window)
• Express Gateway (opens new window)
• Grafana (opens new window)
• Hub'eau (opens new window)
• k2 (opens new window)
• kano (opens new window)
• kapture (opens new window)
• MapServer (opens new window)
• maputnik (opens new window)
• Metar/Taf (opens new window)
• Meteo Radar (opens new window)
• MongoDB (opens new window)
• Mongo Express (opens new window)
• OpenAQ (opens new window)
• OpenRadiation (opens new window)
• PostGis (opens new window)
• Prometheus (opens new window)
• Logspout (opens new window)
• redis (opens new window)
• Redis Commander (opens new window)
• Teleray (opens new window)
• TileServer-GL (opens new window)
• Thredds (opens new window)
• Vigicrues (opens new window)
• Weacast API (opens new window)
• Weacast loaders (opens new window)

TIP

Redis is required to enable Kano and Weacast to communicate in Cloud environments using feathers-distributed (opens new window) under-the-hood.

By default, most of the images of the services are pulled from the Docker Hub (opens new window). Meanwhile Kargo provides a mean to build some of them to ensure better customization and maintainability:

• Express Gateway
• Logspout
• TileServer-GL
• MapServer
• Kano
• The Weacast loaders