feat(workers): add release stage badges to compute provider picker

[rossnelson]

Description & motivation 💭

With AWS Lambda going Public Preview and GCP Cloud Run entering Pre-release, the Compute Provider section gave no indication of either provider's release stage. This differs from features like Standalone Activities, which show a Public Preview label. Since users follow the same worker deployment creation flow for both providers, the missing labels could create confusion about the maturity and availability of each option.

This adds a release stage badge next to each Compute Provider option:

• AWS Lambda → Public Preview
• Google Cloud Run → Pre-release

The badge uses the same secondary Holocene Badge that Nexus and SAA use for their pre-release/preview labels, rendered in the existing labelBadge snippet on RadioCard.

Configurable by the caller

The stage is not hardcoded to the provider — ComputeProviderOption gains an optional releaseStage, so Cloud can drive it through the computeProviders prop it already passes into serverless-worker-create, worker-deployment-version-create, and worker-deployment-version-edit. No new prop to thread.

export type ComputeProviderReleaseStage =
  | 'public-preview'
  | 'pre-release'
  | 'generally-available';

• Unset → falls back to defaultReleaseStage in shared.ts (the two stages above), so OSS/self-hosted and any caller that doesn't set it still get correct labels, and Cloud picks up stage changes on its next UI bump with no Cloud-side change.
• 'generally-available' → renders no badge. This is how a provider drops the label once it GAs. Note undefined means "use the default", never "no badge".
• A disabled option still shows its disabledReason badge ("Coming Soon", "Not supported in AWS namespaces") instead — that's the more relevant maturity signal, and it avoids stacking two badges.

An enum rather than a free-form string keeps the label text and its translation in this repo; Cloud picks a stage, not a string to display. defaultReleaseStage is exported so Cloud can read the defaults if it needs to reason about them.

Screenshots (if applicable) 📸

Design Considerations 🎨

Matches the crude mockup from the thread — badge inline to the right of the provider name. Reused the existing secondary badge type rather than introducing a release-stage-specific style, for consistency with the Pre-release badge already on the Nexus Operations page.

Testing 🧪

How was this tested 👻

• Manual testing
• E2E tests added
• Unit tests added

Storybook stories in compute-provider-picker.stories.svelte:

• "Both enabled (cross-cloud)" now asserts both Public Preview and Pre-release render.
• New "Release stage overridden by caller" story asserts a caller-supplied 'generally-available' on Lambda renders no badge while Cloud Run keeps Pre-release.

pnpm lint clean (0 errors), pnpm test --run 2573 passed / 183 files. pnpm check is unchanged from main — it reports one pre-existing error (Cannot find module './$types' in the history events route) unrelated to these files.

Steps for others to test: 🚶🏽‍♂️🚶🏽‍♀️

1. Go to Create Worker Deployment and look at the Compute Provider section.
2. AWS Lambda shows Public Preview; Google Cloud Run shows Pre-release.
3. In an environment without the serverScaledProviderCloudRun capability, Cloud Run stays disabled and shows Coming Soon rather than a stage badge.
4. In Storybook: Workers/Compute Provider Picker → the two stories above.

Checklists

Merge Checklist

• Add screenshot to the PR description

Issue(s) closed

Docs

Any docs updates needed?

None.

[vercel]

The latest updates on your projects. Learn more about Vercel for GitHub.

Project	Deployment	Actions	Updated (UTC)
holocene	Ready	Preview	Jul 31, 2026 6:27pm