[DOCS] Error reference still points 5xx and 529 failures to status.claude.com instead of the active gateway or provider

Documentation Type

Incorrect/outdated documentation

Documentation Location

https://code.claude.com/docs/en/errors

Section/Topic

"Server errors" section, especially "API Error: 500 Internal server error" and "API Error: Repeated 529 Overloaded errors"

Current Documentation

The docs currently say:

Claude Code shows the raw API response body for any 5xx status. The example below shows a 500 response: ``text API Error: 500 {"type":"error","error":{"type":"api_error","message":"Internal server error"}} · check status.claude.com `` What to do: * Check status.claude.com for active incidents

And for 529s:

``text API Error: Repeated 529 Overloaded errors · check status.claude.com `` What to do: * Check status.claude.com for capacity notices

What's Wrong or Missing?

This page is outdated for the behavior shipped in v2.1.143.

The changelog for v2.1.143 says Claude Code fixed 5xx error messages that were pointing users at status.claude.com instead of naming the configured gateway or cloud provider.

The error reference still documents the older first-party-only wording for both the 500 example and the repeated 529 example. That leaves third-party and gateway users with instructions that imply Anthropic's status page is always the right place to check, even when the failing backend is Bedrock, Vertex, Foundry, or a configured gateway.

This is one root cause: the error reference has not been updated to match the provider-aware 5xx/529 messaging now used by Claude Code.

Suggested Improvement

Update the errors page so the documented examples and troubleshooting guidance match the provider-aware behavior introduced in v2.1.143.

Suggested changes:

1. Replace the hardcoded check status.claude.com examples for 500 and repeated 529 errors with wording that reflects the active backend.
2. Clarify that Anthropic's status page is appropriate for Anthropic-operated services, while third-party or gateway deployments should check the named provider or configured gateway instead.
3. If the exact message varies by backend, document the pattern rather than one fixed string.

For example, the page could describe the behavior like this:

On Anthropic-operated services, Claude Code points you to status.claude.com. On gateway or cloud-provider deployments, Claude Code names the configured gateway or provider so you can check the correct health surface.

Impact

Medium - Makes feature difficult to understand

Additional Context

Affected Pages:

| Page | Line(s) | Context |
|------|---------|---------|
| https://code.claude.com/docs/en/errors | 70-98 | 500 and repeated 529 examples still hardcode status.claude.com and tell all users to check Anthropic's status page |

Total scope: 1 page affected

Relevant changelog context: v2.1.143 says, "Fixed 5xx error messages pointing at status.claude.com instead of naming the configured gateway or cloud provider."