Intelligent Voice recommends a fully-automated pipeline to deploy container images, and a staged deployment pipeline to update to new versions.
IV container images should be mirrored in your own local Container Registry by your Continuous Integration (CI) pipeline, and IV models and datasets should be mirrored in your own local storage. This removes the need for your production VMs to make outbound connections.
Configuration files should be versioned in your local repository (e.g. GitHub or Azure Repos) and environment variables substituted by your Continuous Delivery (CD) pipeline.
IV supplies an Ansible-based installer to provision Docker-enabled Virtual Machines. These can be imaged for auto-scaling. Deploying IV images to Container Instances is also supported.
IV models and datasets should be deployed to persistent storage such as Azure Files.