[Automated] Draft docs (agentgateway): health: only consider server-side grpc errors as unhealthy

[github-actions]

agentgateway/agentgateway#3897 — health: only consider server-side grpc errors as unhealthy

Docs versions: standalone/main, kubernetes/main

No code-path rule covers 1 touched file(s). The planner could route these only by identifier search, so the candidate pages may miss what they changed, including in the other mode:

• crates/agentgateway/src/http/health.rs

What changes for users: The default health classifier no longer treats client-side gRPC statuses, such as INVALID_ARGUMENT, as unhealthy. Server-side gRPC failures, such as INTERNAL, still lower health and can trigger default failover eviction.

What the docs now say: The standalone and Kubernetes virtual-model failover pages now say that default eviction covers server-side gRPC failures, not every non-zero gRPC status.

Not verified: No cluster or local agentgateway run; the configuration is unrun.

How agentgateway-3897 was drafted, and what was not verified

• Code PR: health: only consider server-side grpc errors as unhealthy agentgateway#3897
• Code issue: none linked
• Product: agentgateway (upstream) → agentgateway/website
• Docs version: standalone/main and kubernetes/main
• Category: scoped-edit / pr

Plan, and what changed it

• Planned: Check the observability and access-log targets that the code-path rules named.
• Learned: the code diff and the health.rs unrouted file showed a health-classification change, and the main virtual-model pages carried the stale failover wording.
• Did: Updated the standalone and Kubernetes virtual-model failover pages and wrote a release-note entry.

Why a documentation change is necessary

The pull request changes the default health classifier from any non-zero gRPC status to only server-side gRPC failures. The standalone virtual-model page said that non-zero gRPC statuses were unhealthy, which became wrong. The Kubernetes virtual-model page documented default failover eviction for 5xx responses and connection failures, and needed the server-side gRPC failure case added for the same runtime behavior.

What changed on disk

• content/docs/standalone/main/documentation/llm/virtual-models.md: Replaced the non-zero gRPC status wording with server-side gRPC failures in the failover health table and unhealthyExpression bullet.
• content/docs/kubernetes/main/documentation/llm/models/virtual.md: Added server-side gRPC failures to the default failover eviction description.

How to verify this change

Cannot be verified from the diff alone. The dossier includes unit-test assertions for status 13 as unhealthy and status 3 as healthy, but it does not include a runnable documentation setup that produces those provider responses.

What was not verified

• The configuration was not applied to a cluster or run locally, so the documented behavior is unrun.
• The docs were not checked against a live gRPC provider that returns client-side and server-side gRPC statuses.

How the pages were chosen (triage report)

Draft a documentation change

agentgateway/agentgateway#3897 — health: only consider server-side grpc errors as unhealthy

• Product: agentgateway (upstream) → agentgateway/website
• Docs version: standalone/main, kubernetes/main
• Summary (read from the body): This matches us treating 5xx but not 4xx
• Not tested: no cluster was available, so every command in this diff is unrun
• No rule covers 1 touched file(s). The candidate pages below come from the files that DID match a rule, so they may scope this narrower than the change is. Read these before deciding the change is only about the pages listed:
  • crates/agentgateway/src/http/health.rs

Release notes

This pull request also adds its release-note entries:

• content/docs/kubernetes/main/release-notes/release-notes.md: health: only consider server-side grpc errors as unhealthy (fix, under LLM)
• content/docs/standalone/main/release-notes/release-notes.md: health: only consider server-side grpc errors as unhealthy (fix, under LLM)

Draft branch: pr-tracker-draft-agentgateway-3897

[cloudflare-workers-and-pages]

Deploying agentproxy with    Cloudflare Pages

Latest commit:	500f548
Status:	✅  Deploy successful!
Preview URL:	https://6552ebe9.agentproxy.pages.dev
Branch Preview URL:	https://pr-tracker-draft-agentgatewa-jx8h.agentproxy.pages.dev

View logs