Replace GetProxy with ServeHttp in instance

pkg/instance/instance.go

@@ -5,7 +5,7 @@ import (
 	"fmt"
 	"llamactl/pkg/config"
 	"log"
-	"net/http/httputil"
+	"net/http"
 	"time"
 )
 
@@ -182,15 +182,6 @@ func (i *Instance) GetPort() int {
 	return i.options.GetPort()
 }
 
-// GetProxy returns the reverse proxy for this instance
-func (i *Instance) GetProxy() (*httputil.ReverseProxy, error) {
-	if i.proxy == nil {
-		return nil, fmt.Errorf("instance %s has no proxy component", i.Name)
-	}
-
-	return i.proxy.get()
-}
-
 func (i *Instance) IsRemote() bool {
 	opts := i.GetOptions()
 	if opts == nil {
@@ -242,6 +233,22 @@ func (i *Instance) ShouldTimeout() bool {
 	return i.proxy.shouldTimeout()
 }
 
+// GetInflightRequests returns the current number of inflight requests
+func (i *Instance) GetInflightRequests() int32 {
+	if i.proxy == nil {
+		return 0
+	}
+	return i.proxy.getInflightRequests()
+}
+
+// ServeHTTP serves HTTP requests through the proxy with request tracking and shutdown handling
+func (i *Instance) ServeHTTP(w http.ResponseWriter, r *http.Request) error {
+	if i.proxy == nil {
+		return fmt.Errorf("instance %s has no proxy component", i.Name)
+	}
+	return i.proxy.serveHTTP(w, r)
+}
+
 func (i *Instance) getCommand() string {
 	opts := i.GetOptions()
 	if opts == nil {

pkg/instance/instance_test.go

@@ -171,64 +171,6 @@ func TestSetOptions(t *testing.T) {
 	}
 }
 
-func TestGetProxy(t *testing.T) {
-	globalConfig := &config.AppConfig{
-		Backends: config.BackendConfig{
-			LlamaCpp: config.BackendSettings{
-				Command: "llama-server",
-				Args:    []string{},
-			},
-			MLX: config.BackendSettings{
-				Command: "mlx_lm.server",
-				Args:    []string{},
-			},
-			VLLM: config.BackendSettings{
-				Command: "vllm",
-				Args:    []string{"serve"},
-			},
-		},
-		Instances: config.InstancesConfig{
-			LogsDir: "/tmp/test",
-		},
-		Nodes:     map[string]config.NodeConfig{},
-		LocalNode: "main",
-	}
-
-	options := &instance.Options{
-		Nodes: map[string]struct{}{"main": {}},
-		BackendOptions: backends.Options{
-			BackendType: backends.BackendTypeLlamaCpp,
-			LlamaServerOptions: &backends.LlamaServerOptions{
-				Host: "localhost",
-				Port: 8080,
-			},
-		},
-	}
-
-	// Mock onStatusChange function
-	mockOnStatusChange := func(oldStatus, newStatus instance.Status) {}
-
-	inst := instance.New("test-instance", globalConfig, options, mockOnStatusChange)
-
-	// Get proxy for the first time
-	proxy1, err := inst.GetProxy()
-	if err != nil {
-		t.Fatalf("GetProxy failed: %v", err)
-	}
-	if proxy1 == nil {
-		t.Error("Expected proxy to be created")
-	}
-
-	// Get proxy again - should return cached version
-	proxy2, err := inst.GetProxy()
-	if err != nil {
-		t.Fatalf("GetProxy failed: %v", err)
-	}
-	if proxy1 != proxy2 {
-		t.Error("Expected cached proxy to be returned")
-	}
-}
-
 func TestMarshalJSON(t *testing.T) {
 	globalConfig := &config.AppConfig{
 		Backends: config.BackendConfig{
@@ -613,11 +555,6 @@ func TestRemoteInstanceOperations(t *testing.T) {
 		t.Error("Expected error when restarting remote instance")
 	}
 
-	// GetProxy should fail for remote instance
-	if _, err := inst.GetProxy(); err != nil {
-		t.Error("Expected no error when getting proxy for remote instance")
-	}
-
 	// GetLogs should fail for remote instance
 	if _, err := inst.GetLogs(10); err == nil {
 		t.Error("Expected error when getting logs for remote instance")