From cc7612bdbd622d4504321ff3af3f4d04e929640a Mon Sep 17 00:00:00 2001
From: Elsa <zeyugao@outlook.com>
Date: Thu, 25 Jun 2026 21:17:13 +0800
Subject: [PATCH] Detect OpenAI overloaded error codes

---
 .../service/openai_gateway_service.go         | 15 ++++++++
 ...nai_gateway_service_codex_cli_only_test.go | 12 ++++++
 .../service/openai_gateway_service_test.go    | 38 +++++++++++++++++++
 3 files changed, 65 insertions(+)

diff --git a/backend/internal/service/openai_gateway_service.go b/backend/internal/service/openai_gateway_service.go
index 4ba126e9a3..7f54d49d01 100644
--- a/backend/internal/service/openai_gateway_service.go
+++ b/backend/internal/service/openai_gateway_service.go
@@ -1144,6 +1144,21 @@ func isOpenAIInstructionsRequiredError(upstreamStatusCode int, upstreamMsg strin
 }
 
 func isOpenAITransientProcessingError(upstreamStatusCode int, upstreamMsg string, upstreamBody []byte) bool {
+	if upstreamStatusCode != http.StatusBadRequest && upstreamStatusCode != http.StatusServiceUnavailable {
+		return false
+	}
+
+	hasOpenAIServerOverloadedCode := func(payload []byte) bool {
+		code := strings.ToLower(strings.TrimSpace(gjson.GetBytes(payload, "error.code").String()))
+		if code == "" {
+			code = strings.ToLower(strings.TrimSpace(gjson.GetBytes(payload, "response.error.code").String()))
+		}
+		return code == "server_is_overloaded" || code == "slow_down"
+	}
+
+	if len(upstreamBody) > 0 && hasOpenAIServerOverloadedCode(upstreamBody) {
+		return true
+	}
 	if upstreamStatusCode != http.StatusBadRequest {
 		return false
 	}
diff --git a/backend/internal/service/openai_gateway_service_codex_cli_only_test.go b/backend/internal/service/openai_gateway_service_codex_cli_only_test.go
index 10d5865409..e245e89c0c 100644
--- a/backend/internal/service/openai_gateway_service_codex_cli_only_test.go
+++ b/backend/internal/service/openai_gateway_service_codex_cli_only_test.go
@@ -228,6 +228,18 @@ func TestIsOpenAITransientProcessingError(t *testing.T) {
 		[]byte(`{"error":{"message":"Selected model is at capacity. Please try a different model.","type":"invalid_request_error"}}`),
 	))
 
+	require.True(t, isOpenAITransientProcessingError(
+		http.StatusBadRequest,
+		"",
+		[]byte(`{"error":{"code":"server_is_overloaded","message":"Please retry later.","type":"invalid_request_error"}}`),
+	))
+
+	require.True(t, isOpenAITransientProcessingError(
+		http.StatusServiceUnavailable,
+		"",
+		[]byte(`{"error":{"code":"slow_down","message":"Please retry later."}}`),
+	))
+
 	require.True(t, isOpenAITransientProcessingError(
 		http.StatusBadRequest,
 		"",
diff --git a/backend/internal/service/openai_gateway_service_test.go b/backend/internal/service/openai_gateway_service_test.go
index 51c528cb88..66cfddea6f 100644
--- a/backend/internal/service/openai_gateway_service_test.go
+++ b/backend/internal/service/openai_gateway_service_test.go
@@ -1330,6 +1330,44 @@ func TestOpenAIStreamingResponseFailedBeforeOutputCapacityErrorReturnsFailover(t
 	require.Empty(t, rec.Body.String())
 }
 
+func TestOpenAIStreamingResponseFailedBeforeOutputServerOverloadedCodeReturnsFailover(t *testing.T) {
+	gin.SetMode(gin.TestMode)
+	cfg := &config.Config{
+		Gateway: config.GatewayConfig{
+			StreamDataIntervalTimeout: 0,
+			StreamKeepaliveInterval:   0,
+			MaxLineSize:               defaultMaxLineSize,
+		},
+	}
+	svc := &OpenAIGatewayService{cfg: cfg}
+
+	rec := httptest.NewRecorder()
+	c, _ := gin.CreateTestContext(rec)
+	c.Request = httptest.NewRequest(http.MethodPost, "/", nil)
+
+	resp := &http.Response{
+		StatusCode: http.StatusOK,
+		Body: io.NopCloser(strings.NewReader(strings.Join([]string{
+			"event: response.created",
+			`data: {"type":"response.created","response":{"id":"resp_1"}}`,
+			"",
+			"event: response.failed",
+			`data: {"type":"response.failed","response":{"id":"resp_1","error":{"code":"server_is_overloaded","message":"Please retry later."}}}`,
+			"",
+		}, "\n"))),
+		Header: http.Header{"X-Request-Id": []string{"rid-overloaded-failed"}},
+	}
+
+	_, err := svc.handleStreamingResponse(c.Request.Context(), resp, c, &Account{ID: 1, Platform: PlatformOpenAI, Name: "acc"}, time.Now(), "model", "model")
+	require.Error(t, err)
+	var failoverErr *UpstreamFailoverError
+	require.ErrorAs(t, err, &failoverErr)
+	require.Equal(t, http.StatusBadGateway, failoverErr.StatusCode)
+	require.Contains(t, string(failoverErr.ResponseBody), "Please retry later")
+	require.False(t, c.Writer.Written())
+	require.Empty(t, rec.Body.String())
+}
+
 func TestOpenAIStreamingPreambleOnlyMissingTerminalReturnsFailover(t *testing.T) {
 	gin.SetMode(gin.TestMode)
 	cfg := &config.Config{