Project Layout¶

The repository is organized around the public source tree for Productive K3S profiles and scenarios.

Top-level structure¶

productive-k3s-profiles/
  profiles/
    cloud/
    edge/
    local/
  scenarios/
    cloud/
      aws-single-node/
    edge/
      onprem-basic/
      onprem-basic-arm/
    local/
      multipass/
  docs/
  scripts/

Responsibility split¶

profiles/: public profile defaults and package metadata sidecars
scenarios/: public scenario implementations, helper scripts, and scenario-local Makefiles
docs/: public documentation site for profiles and scenarios
scripts/: repository-local helper scripts such as docs wrappers

Profile defaults and package metadata must stay consistent. Any required input declared as source: package-default in a *.package.yaml sidecar must have a non-empty value in the matching .env file, because that .env becomes the package-level default contract. Installation-specific values should be marked as source: local-override and supplied by the operator through --env-file.

Packaged profiles also declare their package-local scenario path and execution targets in profile.yaml during release generation. Infra consumes that metadata instead of hardcoding scenario names from this repository.

Generated artifacts¶

Each scenario may write generated metadata under its own generated/ directory, typically including things like:

cluster.json
hosts.yml
provider-specific local state or rendered inputs

These artifacts are part of the source-oriented workflow because they expose the resolved runtime view of the scenario.

Notes¶

Note

Public package execution does not happen in this repository. Operators consume the published artifacts through pk3s or productive-k3s-infra.

Note

Compatibility with the Infra engine is validated by engine-side CI cloning this repository into a temporary workspace.