Changeset 0.28.8 (#194)

server/handlers/admin.go

@@ -2,6 +2,7 @@ package handlers
 
 import (
 	"encoding/json"
+	"errors"
 	"fmt"
 	"log"
 	"net/http"
@@ -665,6 +666,10 @@ func (h *AdminHandler) FetchModels(c *gin.Context) {
 
 	added, updated, fetched, err := h.fetchModelsForProvider(c, cfg)
 	if err != nil {
+		if errors.Is(err, ErrUpstreamTimeout) {
+			c.JSON(http.StatusGatewayTimeout, gin.H{"error": err.Error()})
+			return
+		}
 		c.JSON(http.StatusBadGateway, gin.H{"error": "failed to fetch models: " + err.Error()})
 		return
 	}

server/handlers/apiconfigs.go

@@ -1,6 +1,7 @@
 package handlers
 
 import (
+	"errors"
 	"log"
 	"net/http"
 
@@ -279,6 +280,10 @@ func (h *ProviderConfigHandler) FetchModels(c *gin.Context) {
 
 	result, err := syncAndEnableProviderModels(c.Request.Context(), h.stores, cfg, apiKey)
 	if err != nil {
+		if errors.Is(err, ErrUpstreamTimeout) {
+			c.JSON(http.StatusGatewayTimeout, gin.H{"error": err.Error()})
+			return
+		}
 		c.JSON(http.StatusBadGateway, gin.H{"error": "failed to fetch models: " + err.Error()})
 		return
 	}

server/handlers/model_sync.go

@@ -3,14 +3,35 @@ package handlers
 import (
 	"context"
 	"encoding/json"
+	"errors"
 	"fmt"
 	"log"
+	"os"
+	"strconv"
+	"time"
 
 	"git.gobha.me/xcaliber/chat-switchboard/models"
 	"git.gobha.me/xcaliber/chat-switchboard/providers"
 	"git.gobha.me/xcaliber/chat-switchboard/store"
 )
 
+// providerSyncTimeout is the maximum time allowed for an outbound provider
+// HTTP call during model sync. Configurable via PROVIDER_SYNC_TIMEOUT env
+// var (seconds). Default 30s.
+var providerSyncTimeout = func() time.Duration {
+	if v := os.Getenv("PROVIDER_SYNC_TIMEOUT"); v != "" {
+		if n, err := strconv.Atoi(v); err == nil && n > 0 {
+			return time.Duration(n) * time.Second
+		}
+	}
+	return 30 * time.Second
+}()
+
+// ErrUpstreamTimeout is returned when a provider API call exceeds the
+// configured timeout. Handlers use this to distinguish upstream timeouts
+// from other errors and return 504 instead of 502.
+var ErrUpstreamTimeout = errors.New("upstream timeout")
+
 // syncResult holds the outcome of a model sync operation.
 type syncResult struct {
 	Added   int `json:"added"`
@@ -62,8 +83,17 @@ func syncProviderModels(ctx context.Context, stores store.Stores, cfg *models.Pr
 		}
 	}
 
-	provModels, err := prov.ListModels(ctx, provCfg)
+	// Wrap context with timeout for the outbound provider call.
+	// All provider ListModels implementations use http.NewRequestWithContext,
+	// so the deadline propagates to the HTTP transport automatically.
+	syncCtx, cancel := context.WithTimeout(ctx, providerSyncTimeout)
+	defer cancel()
+
+	provModels, err := prov.ListModels(syncCtx, provCfg)
 	if err != nil {
+		if errors.Is(err, context.DeadlineExceeded) {
+			return syncResult{}, fmt.Errorf("%w: %s", ErrUpstreamTimeout, cfg.Provider)
+		}
 		return syncResult{}, err
 	}