What needs documentation?
Add endpoint publishing and verification guidance to the AI Gateway and Connector Gateway deployment guides. Both guides assume operators have reachable gateway URLs, but neither completes the steps to publish them.
- Connector Gateway: the guide sets
authServerIssuer and identifies {issuer}/gw/mcp, but does not explain how to route that hostname to the gateway's data-plane Service. Its REST control plane uses a separate Service and port.
- AI Gateway: the guide covers operator enablement and applying an
AIGateway, but does not explain how to publish and verify the model inference endpoint or distinguish it from the management API.
- Platform deployment: "Expose the platform endpoints" covers the console, Enterprise Manager, and Registry Server. It does not provide gateway publishing instructions.
- Client rollout: "Check network reachability" describes client network requirements but assumes the endpoints are already published.
Add a "Publish and verify the endpoints" section to each gateway deployment guide, covering:
- The client-facing routing target, namespace, ports, and how to discover the installed resources.
- DNS, TLS, hostname, and path requirements, including the Connector Gateway's authorization-server routes and MCP endpoint.
- A source-verified publishing example using a supported ingress or Gateway API approach, including streaming requirements where applicable.
- The separate management/control-plane API target. Explain when it can remain cluster-internal and how to route it if external access is required.
- Verification of endpoint reachability and authenticated use, followed by links to client rollout and the API reference.
Context and references
Current pages:
Related: #1159 tracks the broader Connector Gateway deployment contract; #1175 tracks Connector Gateway documentation improvements. This issue focuses on publishing and verifying endpoints across both gateways.
Use case
As a platform operator, I need to turn an installed gateway into a reachable client endpoint and know which Service each URL targets before distributing client setup instructions.
Additional context
This gap surfaced during API reference discoverability work. That work adds dedicated reference pages and base URL guidance; a complete deployment publishing workflow is a separate follow-up.
What needs documentation?
Add endpoint publishing and verification guidance to the AI Gateway and Connector Gateway deployment guides. Both guides assume operators have reachable gateway URLs, but neither completes the steps to publish them.
authServerIssuerand identifies{issuer}/gw/mcp, but does not explain how to route that hostname to the gateway's data-plane Service. Its REST control plane uses a separate Service and port.AIGateway, but does not explain how to publish and verify the model inference endpoint or distinguish it from the management API.Add a "Publish and verify the endpoints" section to each gateway deployment guide, covering:
Context and references
Current pages:
Related: #1159 tracks the broader Connector Gateway deployment contract; #1175 tracks Connector Gateway documentation improvements. This issue focuses on publishing and verifying endpoints across both gateways.
Use case
As a platform operator, I need to turn an installed gateway into a reachable client endpoint and know which Service each URL targets before distributing client setup instructions.
Additional context
This gap surfaced during API reference discoverability work. That work adds dedicated reference pages and base URL guidance; a complete deployment publishing workflow is a separate follow-up.