AMP Clusters deployment model
The deployment is a single operation, which leaves no way to customize the cluster before the core services are deployed, or even to use an existing swarm cluster.
The worker nodes that are created for the user services (everything that will be deployed by a user) are all of the same type (although it’s a parameter).
Even though it’s a Docker swarm, all docker operations (stack deploy, logs, stats, etc) are done with the AMP CLI.
The deployment is done in a few steps:
- Creation of the Swarm cluster
- Deployment of AMP core services
- Creation of user worker nodes (as many groups as required)
The AMP CLI is only used for the non-native Docker operations (cluster deployment, authentication, etc).
All Docker native operations (stack deploy, service logs, etc) are done with the Docker CLI with mutual auth.
