diff --git a/pkg/kubelet/userns/userns_manager.go b/pkg/kubelet/userns/userns_manager.go
index 56a9a8e8ade..2024c071715 100644
--- a/pkg/kubelet/userns/userns_manager.go
+++ b/pkg/kubelet/userns/userns_manager.go
@@ -55,10 +55,9 @@ type userNsPodsManager interface {
 }
 
 type UsernsManager struct {
-	used         *allocator.AllocationBitmap
-	usedBy       map[types.UID]uint32 // Map pod.UID to range used
-	removed      int
-	numAllocated int
+	used    *allocator.AllocationBitmap
+	usedBy  map[types.UID]uint32 // Map pod.UID to range used
+	removed int
 
 	off int
 	len int
@@ -216,16 +215,6 @@ func (m *UsernsManager) isSet(v uint32) bool {
 // The first return value is the first ID in the user namespace, the second returns
 // the length for the user namespace range.
 func (m *UsernsManager) allocateOne(pod types.UID) (firstID uint32, length uint32, err error) {
-	if m.numAllocated >= maxPods {
-		return 0, 0, fmt.Errorf("limit on count of pods with user namespaces exceeded (limit is %v, current pods with userns: %v)", maxPods, m.numAllocated)
-	}
-	m.numAllocated++
-	defer func() {
-		if err != nil {
-			m.numAllocated--
-		}
-	}()
-
 	firstZero, found, err := m.used.AllocateNext()
 	if err != nil {
 		return 0, 0, err
@@ -265,15 +254,6 @@ func (m *UsernsManager) record(pod types.UID, from, length uint32) (err error) {
 	if found && prevFrom == from {
 		return nil
 	}
-	if m.numAllocated >= maxPods {
-		return fmt.Errorf("limit on count of pods with user namespaces exceeded (limit is %v, current pods with userns: %v)", maxPods, m.numAllocated)
-	}
-	m.numAllocated++
-	defer func() {
-		if err != nil {
-			m.numAllocated--
-		}
-	}()
 
 	klog.V(5).InfoS("new pod user namespace allocation", "podUID", pod)
 
@@ -318,7 +298,6 @@ func (m *UsernsManager) releaseWithLock(pod types.UID) {
 	delete(m.usedBy, pod)
 
 	klog.V(5).InfoS("releasing pod user namespace allocation", "podUID", pod)
-	m.numAllocated--
 	m.removed++
 
 	_ = os.Remove(filepath.Join(m.kl.GetPodDir(pod), mappingsFile))
diff --git a/pkg/kubelet/userns/userns_manager_test.go b/pkg/kubelet/userns/userns_manager_test.go
index 56cb17bda60..db325a2c1e5 100644
--- a/pkg/kubelet/userns/userns_manager_test.go
+++ b/pkg/kubelet/userns/userns_manager_test.go
@@ -378,42 +378,6 @@ func TestCleanupOrphanedPodUsernsAllocations(t *testing.T) {
 	}
 }
 
-func TestAllocateMaxPods(t *testing.T) {
-	defer featuregatetesting.SetFeatureGateDuringTest(t, utilfeature.DefaultFeatureGate, pkgfeatures.UserNamespacesSupport, true)()
-
-	testUserNsPodsManager := &testUserNsPodsManager{}
-	m, err := MakeUserNsManager(testUserNsPodsManager)
-	require.NoError(t, err)
-
-	// The first maxPods allocations should succeed.
-	for i := 0; i < maxPods; i++ {
-		_, _, err = m.allocateOne(types.UID(fmt.Sprintf("%d", i)))
-		require.NoError(t, err)
-	}
-
-	// The next allocation should fail, hitting maxPods.
-	_, _, err = m.allocateOne(types.UID(fmt.Sprintf("%d", maxPods+1)))
-	assert.Error(t, err)
-}
-
-func TestRecordMaxPods(t *testing.T) {
-	defer featuregatetesting.SetFeatureGateDuringTest(t, utilfeature.DefaultFeatureGate, pkgfeatures.UserNamespacesSupport, true)()
-
-	testUserNsPodsManager := &testUserNsPodsManager{}
-	m, err := MakeUserNsManager(testUserNsPodsManager)
-	require.NoError(t, err)
-
-	// The first maxPods allocations should succeed.
-	for i := 0; i < maxPods; i++ {
-		err = m.record(types.UID(fmt.Sprintf("%d", i)), uint32((i+1)*65536), 65536)
-		require.NoError(t, err)
-	}
-
-	// The next allocation should fail, hitting maxPods.
-	err = m.record(types.UID(fmt.Sprintf("%d", maxPods+1)), uint32((maxPods+1)*65536), 65536)
-	assert.Error(t, err)
-}
-
 type failingUserNsPodsManager struct {
 	testUserNsPodsManager
 }