feat(helm/gha-runner-scale-set-experimental): add $extra field passthrough for runner container in dind and kubernetes modes

[stokkie90]

Summary

• In dind and kubernetes modes, the runner container template now reads runner.container as a map and passes any fields not explicitly rendered (securityContext, resources, imagePullPolicy, args, etc.) through to the container spec via the $extra = omit ... | toYaml pattern.
• This mirrors the behaviour already present in mode: "" (empty mode) via _mode_empty.tpl, and in the dind sidecar container in _mode_dind.tpl.
• Input validation (wrong type guards) is included: runner.container must be a map, env/volumeMounts/args must be lists, securityContext must be a map, and volumes is rejected with a clear error pointing to runner.pod.spec.volumes.

Risks & rollout

• Purely additive: existing values that don't set runner.container are unaffected (the map defaults to {}).
• No breaking changes to existing behaviour; the name key in runner.container is silently ignored to preserve the hardcoded runner container name.

Validation

• 14 new helm-unittest test cases added (7 per mode) covering: extra field passthrough, name-override suppression, and all validation failure paths.
• Full test suite passes: 171 tests, 0 failures.

Charts:      1 passed, 1 total
Test Suites: 35 passed, 35 total
Tests:       171 passed, 171 total
Time:        427ms

[Copilot]

Pull request overview

Adds runner.container “extra fields” passthrough for the runner container when using runner.mode: dind and runner.mode: kubernetes, aligning these modes with the existing “mode-empty” behavior and improving configurability of the runner container spec.

Changes:

• Add omit ... | toYaml passthrough for non-explicitly-rendered runner.container fields in dind + kubernetes runner container templates (while ignoring runner.container.name).
• Add type-guard validation for runner.container and selected fields (env, volumeMounts, args, securityContext, and disallow volumes) in both modes.
• Add helm-unittest coverage for passthrough behavior + validation failure paths in both modes.

Reviewed changes

Copilot reviewed 5 out of 5 changed files in this pull request and generated 2 comments.

File	Description
charts/gha-runner-scale-set-experimental/templates/_mode_kubernetes.tpl	Adds validation and $extra passthrough injection for the kubernetes-mode runner container.
charts/gha-runner-scale-set-experimental/templates/_mode_dind.tpl	Adds validation and $extra passthrough injection for the dind-mode runner container.
charts/gha-runner-scale-set-experimental/tests/autoscaling_runner_set_kubernetes_mode_spec_test.yaml	Adds unit tests covering kubernetes-mode passthrough, name-override suppression, and validation errors.
charts/gha-runner-scale-set-experimental/tests/autoscaling_runner_set_dind_mode_spec_test.yaml	Adds unit tests covering dind-mode passthrough, name-override suppression, and validation errors.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[gazab]

We would love to be able to extends env and volumeMounts in the runner container when using mode dind as well!

[stokkie90]

Addressed both Copilot review comments:

1. _mode_dind.tpl — use $runner intermediate variable: fixed by introducing $runner := (.Values.runner | default dict) before accessing runner.container, consistent with the kubernetes mode pattern.

2. Duplicate validation block: extracted the six type-guard checks into a shared runner.container.validate helper in _defaults.tpl. Both _mode_dind.tpl and _mode_kubernetes.tpl now call {{- include "runner.container.validate" . -}} instead of duplicating the block.

All 171 tests continue to pass.