I’ve got a deployment that creates dynamic transient infrastructure into a QA environment - specifically an Azure web apps created in a resource group created during the first step of the deployment. When the apps are created, they’re tagged with a unique identifier to allow us to logically segregate them, and when the targets are created in Octopus, they are assigned to the QA environment with the individual roles “app” and “api” as well as a role for the unique identifier for that deployment so that we don’t have transient environments to manage in Octopus as that would lead down a path that I predict would be messy.
Normally I would segregate infrastructure by role, but there doesn’t appear to be a way to deploy to a variable role. So when I deploy to the “app” role for instance, the deployment happens to all infrastructure in the QA environment tagged with the “app” role. I need to further isolate a target for the deployment step to only the “app” roles for the unique identifier for this deployment.
Can you offer any insight into the approach I should take for this?