From 31b1ed2c965abbe0e79c3b585557f301e0902de4 Mon Sep 17 00:00:00 2001
From: JianFeeeee <root@qq.com>
Date: Mon, 24 Aug 2026 16:02:31 +0800
Subject: [PATCH] feat(gateway): collapse multi-tier failure details into short
 client message

Clients saw the full per-tier chain error including upstream HTML WAF
pages and JSON quota payloads. The response now carries one capped
one-line reason per tier (quota/cooling skips preserved); full detail
remains in rec.Err / stats API and is logged server-side.
---
 internal/gateway/chat.go         | 45 ++++++++++++++++++++++++++++----
 internal/gateway/gateway_test.go |  2 +-
 2 files changed, 41 insertions(+), 6 deletions(-)

diff --git a/internal/gateway/chat.go b/internal/gateway/chat.go
index b2e447e..6c97490 100644
--- a/internal/gateway/chat.go
+++ b/internal/gateway/chat.go
@@ -5,6 +5,7 @@ import (
 	"encoding/json"
 	"errors"
 	"fmt"
+	"log"
 	"net/http"
 	"strings"
 	"sync/atomic"
@@ -523,6 +524,40 @@ func upstreamErrStatus(err error) int {
 	return http.StatusBadGateway
 }
 
+// clientUpstreamErr collapses upstream failure details into a short message
+// for the client: per-tier bodies (WAF HTML pages, quota payloads, ...) stay
+// in rec.Err / the stats API and the server log instead of the response.
+// Per-tier one-line reasons are kept (quota/cooling skips carry no error body
+// and are the actionable part); each is capped so HTML dumps can't leak.
+func clientUpstreamErr(err error) string {
+	log.Printf("[gateway] upstream failure surfaced to client: %v", err)
+	var ce *scheduler.ChainErr
+	if errors.As(err, &ce) {
+		parts := make([]string, 0, len(ce.Tiers)+len(ce.Skipped))
+		for _, t := range ce.Tiers {
+			parts = append(parts, oneLine(fmt.Sprintf("%s/%s: %v", t.Source, t.Model, t.Err), 80))
+		}
+		for _, sk := range ce.Skipped {
+			parts = append(parts, oneLine(sk, 80))
+		}
+		msg := strings.Join(parts, "; ")
+		if len(msg) > 300 {
+			msg = msg[:300] + "..."
+		}
+		return fmt.Sprintf("all %d auto providers failed: %s", len(parts), msg)
+	}
+	return oneLine(err.Error(), 160)
+}
+
+// oneLine flattens an error string to a single line capped at n chars.
+func oneLine(s string, n int) string {
+	s = strings.Join(strings.Fields(s), " ")
+	if len(s) > n {
+		s = s[:n] + "..."
+	}
+	return s
+}
+
 func (g *Gateway) singleChat(w http.ResponseWriter, ctx context.Context, cands []*provider.Provider, req *types.ChatRequest, effective string, rec *Req) {
 	rec.LatMs = 0
 	t0 := time.Now()
@@ -533,7 +568,7 @@ func (g *Gateway) singleChat(w http.ResponseWriter, ctx context.Context, cands [
 		rec.Status = upstreamErrStatus(err)
 		rec.Err = err.Error()
 		g.writeRec(rec)
-		writeError(w, rec.Status, "upstream_error", err.Error())
+		writeError(w, rec.Status, "upstream_error", clientUpstreamErr(err))
 		return
 	}
 	rec.OK = true
@@ -620,7 +655,7 @@ func (g *Gateway) streamChat(w http.ResponseWriter, ctx context.Context, cands [
 		rec.OK = false
 		rec.Status = upstreamErrStatus(err)
 		rec.Err = err.Error()
-		writeError(w, rec.Status, "upstream_error", err.Error())
+		writeError(w, rec.Status, "upstream_error", clientUpstreamErr(err))
 		return
 	}
 	if usedModel != "" {
@@ -741,7 +776,7 @@ func (g *Gateway) singleChatAuto(w http.ResponseWriter, ctx context.Context, cha
 			rec.Model = ce.Tiers[0].Model
 		}
 		g.writeRec(rec)
-		writeError(w, rec.Status, "upstream_error", err.Error())
+		writeError(w, rec.Status, "upstream_error", clientUpstreamErr(err))
 		return
 	}
 	rec.OK = true
@@ -799,7 +834,7 @@ func (g *Gateway) streamChatAuto(w http.ResponseWriter, ctx context.Context, cha
 			rec.Source = ce.Tiers[0].Source
 			rec.Model = ce.Tiers[0].Model
 		}
-		writeError(w, rec.Status, "upstream_error", err.Error())
+		writeError(w, rec.Status, "upstream_error", clientUpstreamErr(err))
 		return
 	}
 	if usedModel != "" {
@@ -946,7 +981,7 @@ func (g *Gateway) handleImage(w http.ResponseWriter, r *http.Request) {
 		rec.Status = upstreamErrStatus(err)
 		rec.Err = err.Error()
 		g.writeRec(rec)
-		writeError(w, rec.Status, "upstream_error", err.Error())
+		writeError(w, rec.Status, "upstream_error", clientUpstreamErr(err))
 		return
 	}
 	if usedSrc != "" {
diff --git a/internal/gateway/gateway_test.go b/internal/gateway/gateway_test.go
index f8e9ba1..b2e161e 100644
--- a/internal/gateway/gateway_test.go
+++ b/internal/gateway/gateway_test.go
@@ -139,7 +139,7 @@ func TestChatAutoChain503Summary(t *testing.T) {
 	if rr.Code != http.StatusServiceUnavailable {
 		t.Fatalf("status=%d body=%s", rr.Code, rr.Body.String())
 	}
-	if !strings.Contains(rr.Body.String(), "all auto tiers failed") ||
+	if !strings.Contains(rr.Body.String(), "auto providers failed") ||
 		!strings.Contains(rr.Body.String(), "a/a-m") ||
 		!strings.Contains(rr.Body.String(), "b/b-m") {
 		t.Fatalf("503 must summarize every tier, body=%s", rr.Body.String())