NOTE: The code sample contains a sample historically used with legacy Jenkins distributed architecture terminology. Such language isn't accepted anymore and should be read as "jenkins_agent" instead, the new service name.
The docker_containers pillar entry contains mixed data, according we need one container or several containers for a service.
Example of one service, two containers:
jenkins_slave: apsile: ip: 172.17.0.100 elapsi: ip: 172.17.0.101
Example of one service, one container:
cachet: app_port: 39080 host: status.nasqueron.org mysql_link: acquisitariat
To make easier to write Salt states depending of this pillar entry, plan is to standardize the configuration entry to one service, several container.
Accordingly, all the containers deployment logic should support several containers for one service.