fix(frontend): meaningful serverless metadata errors

[jog1t]

Description

Please include a summary of the changes and the related issue. Please also include relevant motivation and context.

Type of change

• Bug fix (non-breaking change which fixes an issue)
• New feature (non-breaking change which adds functionality)
• Breaking change (fix or feature that would cause existing functionality to not work as expected)
• This change requires a documentation update

How Has This Been Tested?

Please describe the tests that you ran to verify your changes.

Checklist:

• My code follows the style guidelines of this project
• I have performed a self-review of my code
• I have commented my code, particularly in hard-to-understand areas
• I have made corresponding changes to the documentation
• My changes generate no new warnings
• I have added tests that prove my fix is effective or that my feature works
• New and existing unit tests pass locally with my changes

[railway-app]

🚅 Deployed to the rivet-pr-5102 environment in rivet-frontend

Service	Status	Web	Updated (UTC)
kitchen-sink	❌ Build Failed (View Logs)	Web	May 26, 2026 at 10:24 pm
frontend-cloud	🕒 Building (View Logs)	Web	May 26, 2026 at 10:23 pm
ladle	🕒 Building (View Logs)	Web	May 26, 2026 at 10:23 pm
frontend-inspector	🕒 Building (View Logs)	Web	May 26, 2026 at 10:23 pm
website	😴 Sleeping (View Logs)	Web	May 26, 2026 at 8:32 pm
mcp-hub	✅ Success (View Logs)	Web	May 26, 2026 at 8:22 pm

[jog1t]

• refactor(frontend): runner_pool -> rivet_pool #5107
• refactor(frontend): modify onboarding for rivet compute #5106
• fix(frontend): align log region column and disable selection on metadata #5105
• fix(frontend): clickable namespaces in deployments table #5104
• fix(fontend): less monoscaped #5103
• fix(frontend): meaningful serverless metadata errors #5102 👈 (View in Graphite)
• main

This stack of pull requests is managed by Graphite. Learn more about stacking.

[claude]

Code Review

Summary: This PR correctly fixes the root cause. The old formatMetadataError was pattern-matching against camelCase TypeScript-generated discriminated-union keys (invalidRequest, nonSuccessStatus, etc.) while the actual wire format sends a message/details/metadata envelope with snake_case metadata.kind. Every variant silently fell through to 'Unknown error'. The new approach is more robust, surfaces the server human-readable message field directly, and degrades gracefully for unknown future variants.

---

1. Untruncated server body in UI error string (moderate)

serverless-health-error.ts L49 appends metadata.body for non_success_status (and falls back to it for invalid_response_json when parse_error is absent) without any length cap. When a WAF, CDN, or misbehaving upstream returns a full HTML page as the response body, the UI renders the entire raw HTML inline as the error string. The existing network-error path in extractErrorMessage already applies .slice(0, 200) for this reason; the same cap should apply to any externally-sourced field appended by formatServerlessMetadataError.

2. Redundant .partial() on an already-optional schema (minor)

serverless-health-error.ts L6-L17: all fields already carry .optional(), making the trailing .partial() call a no-op in Zod. Drop .partial() or the per-field .optional() calls to clarify intent.

---

Positive notes: schema-based parsing with z.unknown() is the right fix; the forward-compat design (unknown kind -> surface message -> humanise kind -> generic fallback) is well tested; stories provide a clear visual audit trail; extracting the formatter into its own file improves testability.

[jog1t]

Merge activity

• May 26, 10:23 PM UTC: A user started a stack merge that includes this pull request via Graphite.
• May 26, 10:24 PM UTC: Graphite rebased this pull request as part of a merge.
• May 26, 10:25 PM UTC: @jog1t merged this pull request with Graphite.