kubectl: Cache Verifier.HasSupport calls

The underlying implementation decodes OpenAPI schema on every HasSupport
call. This causes these calls to be expensive and noticable when many
resources are being applied.

This patch wraps the verifier with a thread-safe cache so that when many
resources of the same kind are being checked, only the first call
includes schema decoding.

This solution is specifically limited to the kubectl codebase without
any changes required in client-go, for now, although that is where the
issue is actually originating.

Kubernetes-commit: 9043afae6d98585dd14d203d48807b14b433d730

pkg/cmd/util/caching_verifier.go

@@ -0,0 +1,59 @@
+/*
+Copyright 2025 The Kubernetes Authors.
+
+Licensed under the Apache License, Version 2.0 (the "License");
+you may not use this file except in compliance with the License.
+You may obtain a copy of the License at
+
+    http://www.apache.org/licenses/LICENSE-2.0
+
+Unless required by applicable law or agreed to in writing, software
+distributed under the License is distributed on an "AS IS" BASIS,
+WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+See the License for the specific language governing permissions and
+limitations under the License.
+*/
+
+package util
+
+import (
+	"sync"
+
+	"k8s.io/apimachinery/pkg/runtime/schema"
+	"k8s.io/cli-runtime/pkg/resource"
+)
+
+// cachingVerifier wraps the given Verifier to cache return values forever.
+type cachingVerifier struct {
+	cache map[schema.GroupVersionKind]error
+	mu    sync.RWMutex
+	next  resource.Verifier
+}
+
+// newCachingVerifier creates a new cache using the given underlying verifier.
+func newCachingVerifier(next resource.Verifier) *cachingVerifier {
+	return &cachingVerifier{
+		cache: make(map[schema.GroupVersionKind]error),
+		next:  next,
+	}
+}
+
+// HasSupport implements resource.Verifier. It cached return values from the underlying verifier forever.
+func (cv *cachingVerifier) HasSupport(gvk schema.GroupVersionKind) error {
+	// Try to get the cached value.
+	cv.mu.RLock()
+	err, ok := cv.cache[gvk]
+	cv.mu.RUnlock()
+	if ok {
+		return err
+	}
+
+	// Cache miss. Get the actual result.
+	err = cv.next.HasSupport(gvk)
+
+	// Update the cache.
+	cv.mu.Lock()
+	cv.cache[gvk] = err
+	cv.mu.Unlock()
+	return err
+}

pkg/cmd/util/caching_verifier_test.go

@@ -0,0 +1,105 @@
+/*
+Copyright 2025 The Kubernetes Authors.
+
+Licensed under the Apache License, Version 2.0 (the "License");
+you may not use this file except in compliance with the License.
+You may obtain a copy of the License at
+
+    http://www.apache.org/licenses/LICENSE-2.0
+
+Unless required by applicable law or agreed to in writing, software
+distributed under the License is distributed on an "AS IS" BASIS,
+WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+See the License for the specific language governing permissions and
+limitations under the License.
+*/
+
+package util
+
+import (
+	"errors"
+	"testing"
+
+	"k8s.io/apimachinery/pkg/runtime/schema"
+)
+
+type verifierCall struct {
+	GVK schema.GroupVersionKind
+	Err error
+}
+
+type mockVerifier struct {
+	t             *testing.T
+	expectedCalls map[schema.GroupVersionKind]error
+}
+
+func (m *mockVerifier) CheckExpectations() {
+	if len(m.expectedCalls) != 0 {
+		m.t.Errorf("Expected calls remaining: %v", m.expectedCalls)
+	}
+}
+
+func (m *mockVerifier) HasSupport(gvk schema.GroupVersionKind) error {
+	returnErr, ok := m.expectedCalls[gvk]
+	if !ok {
+		m.t.Errorf("Unexpected HasSupport call with GVK=%v", gvk)
+	}
+	delete(m.expectedCalls, gvk)
+	return returnErr
+}
+
+func TestCachingVerifier(t *testing.T) {
+	gvk1 := schema.GroupVersionKind{
+		Group:   "group",
+		Version: "version",
+		Kind:    "kind",
+	}
+	gvk2 := schema.GroupVersionKind{
+		Group:   "group2",
+		Version: "version2",
+		Kind:    "kind2",
+	}
+
+	err1 := errors.New("some error")
+
+	testCases := []struct {
+		name                    string
+		calls                   []verifierCall
+		expectedUnderlyingCalls map[schema.GroupVersionKind]error
+	}{
+		{
+			name: "return value is cached",
+			calls: []verifierCall{
+				{GVK: gvk1, Err: nil},
+				{GVK: gvk1, Err: nil},
+				{GVK: gvk1, Err: nil},
+				{GVK: gvk2, Err: err1},
+				{GVK: gvk2, Err: err1},
+				{GVK: gvk2, Err: err1},
+			},
+			expectedUnderlyingCalls: map[schema.GroupVersionKind]error{
+				gvk1: nil,
+				gvk2: err1,
+			},
+		},
+	}
+
+	for _, tc := range testCases {
+		t.Run(tc.name, func(t *testing.T) {
+			m := &mockVerifier{
+				t:             t,
+				expectedCalls: tc.expectedUnderlyingCalls,
+			}
+			verifier := newCachingVerifier(m)
+
+			for _, call := range tc.calls {
+				err := verifier.HasSupport(call.GVK)
+				if !errors.Is(err, call.Err) {
+					t.Errorf("Expected error: %v, got: %v", call.Err, err)
+				}
+			}
+
+			m.CheckExpectations()
+		})
+	}
+}

pkg/cmd/util/factory_client_access.go

@@ -172,7 +172,7 @@ func (f *factoryImpl) Validator(validationDirective string) (validation.Schema,
 	// the discovery client.
 	oapiV3Client := cached.NewClient(discoveryClient.OpenAPIV3())
 	queryParam := resource.QueryParamFieldValidation
-	primary := resource.NewQueryParamVerifierV3(dynamicClient, oapiV3Client, queryParam)
+	primary := newCachingVerifier(resource.NewQueryParamVerifierV3(dynamicClient, oapiV3Client, queryParam))
 	secondary := resource.NewQueryParamVerifier(dynamicClient, f.openAPIGetter(), queryParam)
 	fallback := resource.NewFallbackQueryParamVerifier(primary, secondary)
 	return validation.NewParamVerifyingSchema(schema, fallback, string(validationDirective)), nil