The deployment plan is created and the event is published.
The Controller reacts on the “new deployment” event, changes the deployment state to “in progress”, and dispatches Workorders from step 1 to Inductors.
The inductors consume the Workorders.
Call provider API instantiates the Compute, Storage, etc.
After the compute is up, there is an ssh remote execution of Chef recipes to deploy software, configuration, monitors, etc.
The inductor sends the Workorder result back to the Controller to update the CMS.
After all Workorders from step 1 are successfully processed, move to the next step.
Metrics and Logs are collected (via flume) to Cassandra.