Record actual Responses endpoints in tracing spans (#40906)

## Why

Responses-compatible requests can use routes other than `/responses`, but their
tracing spans reported the default route regardless of the selected endpoint.

## What changed

- Populate `api.path` from the selected `ResponsesEndpoint` for HTTP and
  WebSocket requests.
- Retain the endpoint on WebSocket connections so stream-request spans report
  the same route as connection spans.

GitOrigin-RevId: 349eca94b27d4cc3a1adde8b257313506bdaf0f4

codex-rs/codex-api/src/endpoint/responses.rs

@@ -96,7 +96,7 @@ impl<T: HttpTransport> ResponsesClient<T> {
         fields(
             transport = "responses_http",
             http.method = "POST",
-            api.path = "responses"
+            api.path = self.endpoint.path()
         )
     )]
     pub async fn stream_request(
@@ -136,7 +136,7 @@ impl<T: HttpTransport> ResponsesClient<T> {
         fields(
             transport = "responses_http",
             http.method = "POST",
-            api.path = "responses",
+            api.path = self.endpoint.path(),
             turn.has_state = turn_state.is_some()
         )
     )]

codex-rs/codex-api/src/endpoint/responses_websocket.rs

@@ -182,6 +182,7 @@ struct ResponsesWebsocketTimingLogContext {
 
 pub struct ResponsesWebsocketConnection {
     stream: Arc<Mutex<Option<WsStream>>>,
+    endpoint: ResponsesEndpoint,
     // TODO (pakrym): is this the right place for timeout?
     idle_timeout: Duration,
     server_reasoning_included: bool,
@@ -193,6 +194,7 @@ impl std::fmt::Debug for ResponsesWebsocketConnection {
     fn fmt(&self, f: &mut std::fmt::Formatter<'_>) -> std::fmt::Result {
         f.debug_struct("ResponsesWebsocketConnection")
             .field("stream", &"<ws-stream>")
+            .field("endpoint", &self.endpoint)
             .field("idle_timeout", &self.idle_timeout)
             .field("server_reasoning_included", &self.server_reasoning_included)
             .field("server_model", &self.server_model)
@@ -208,9 +210,11 @@ impl ResponsesWebsocketConnection {
         server_reasoning_included: bool,
         server_model: Option<String>,
         telemetry: Option<Arc<dyn WebsocketTelemetry>>,
+        endpoint: ResponsesEndpoint,
     ) -> Self {
         Self {
             stream: Arc::new(Mutex::new(Some(stream))),
+            endpoint,
             idle_timeout,
             server_reasoning_included,
             server_model,
@@ -226,7 +230,7 @@ impl ResponsesWebsocketConnection {
         name = "responses_websocket.stream_request",
         level = "info",
         skip_all,
-        fields(transport = "responses_websocket", api.path = "responses")
+        fields(transport = "responses_websocket", api.path = self.endpoint.path())
     )]
     pub async fn stream_request(
         &self,
@@ -388,7 +392,7 @@ impl ResponsesWebsocketClient {
         name = "responses_websocket.connect",
         level = "info",
         skip_all,
-        fields(transport = "responses_websocket", api.path = "responses")
+        fields(transport = "responses_websocket", api.path = self.endpoint.path())
     )]
     pub async fn connect(
         &self,
@@ -415,6 +419,7 @@ impl ResponsesWebsocketClient {
             server_reasoning_included,
             server_model,
             telemetry,
+            self.endpoint,
         ))
     }