Support non-containerized Applications

Overview

Expand Radius Canvas to model and deploy Applications that do not use Dockerfiles or container images. This would extend the integration to workloads such as serverless functions, managed compute services, scripts, and other application components that cannot be represented as  Radius.Compute/containerImages  and  Radius.Compute/containers .

Radius skills should analyze the repository, identify supported workload types, and generate or update a  .radius/app.bicep  application definition. Radius Canvas should display these workloads and their dependencies in the Modeled, Planned, Deployed, and Diff views.

The deployment experience should provision the required cloud resources, package or publish application artifacts when necessary, monitor deployment progress, and support deletion through the existing GitHub Actions workflow model.

Expected outcomes

• Detect supported non-containerized workloads.
• Generate valid Radius resource definitions for those workloads.
• Visualize workloads, infrastructure, dependencies, and source references.
• Plan and deploy the Application through GitHub Actions.
• Display deployment status and resulting cloud resources.
• Delete deployed workloads and associated infrastructure.
• Clearly explain unsupported workload types instead of generating an incomplete model.
• Preserve the existing containerized-application workflow.

Additional context

The current preview models and deploys containerized Applications with Dockerfiles. This roadmap item expands the supported compute models while maintaining a consistent Radius Canvas experience.

Start by reviewing the existing .radius/app.bicep application definition and GitHub Actions workflow model, then inventory the supported non-containerized workload types and their required cloud resources. Define how Radius skills and Canvas views should represent supported and unsupported workloads. Done means these workloads can be modeled, planned, deployed, monitored, and deleted without regressing containerized applications.