diff --git a/metropolis/test/e2e/suites/kubernetes/kubernetes_helpers.go b/metropolis/test/e2e/suites/kubernetes/kubernetes_helpers.go
index 9c67117..c8e04bf 100644
--- a/metropolis/test/e2e/suites/kubernetes/kubernetes_helpers.go
+++ b/metropolis/test/e2e/suites/kubernetes/kubernetes_helpers.go
@@ -73,14 +73,13 @@
 // makeHTTPServerDeploymentSpec generates the deployment spec for the test HTTP
 // server.
 func makeHTTPServerDeploymentSpec(name string) *appsv1.Deployment {
-	oneVal := int32(1)
 	return &appsv1.Deployment{
 		ObjectMeta: metav1.ObjectMeta{Name: name},
 		Spec: appsv1.DeploymentSpec{
 			Selector: &metav1.LabelSelector{MatchLabels: map[string]string{
 				"name": name,
 			}},
-			Replicas: &oneVal,
+			Replicas: ptr.To(int32(1)),
 			Template: corev1.PodTemplateSpec{
 				ObjectMeta: metav1.ObjectMeta{
 					Labels: map[string]string{
@@ -129,11 +128,10 @@
 
 // makeSelftestSpec generates a Job spec for the E2E self-test image.
 func makeSelftestSpec(name string) *batchv1.Job {
-	one := int32(1)
 	return &batchv1.Job{
 		ObjectMeta: metav1.ObjectMeta{Name: name},
 		Spec: batchv1.JobSpec{
-			BackoffLimit: &one,
+			BackoffLimit: ptr.To(int32(1)),
 			Template: corev1.PodTemplateSpec{
 				ObjectMeta: metav1.ObjectMeta{
 					Labels: map[string]string{
diff --git a/metropolis/test/e2e/suites/kubernetes/run_test.go b/metropolis/test/e2e/suites/kubernetes/run_test.go
index 18239e0..c7f5612 100644
--- a/metropolis/test/e2e/suites/kubernetes/run_test.go
+++ b/metropolis/test/e2e/suites/kubernetes/run_test.go
@@ -132,12 +132,11 @@
 	}
 	// Nominate both nodes to be Kubernetes workers.
 	for _, nid := range cluster.NodeIDs {
-		yes := true
 		_, err := mgmt.UpdateNodeRoles(ctx, &apb.UpdateNodeRolesRequest{
 			Node: &apb.UpdateNodeRolesRequest_Id{
 				Id: nid,
 			},
-			KubernetesWorker: &yes,
+			KubernetesWorker: ptr.To(true),
 		})
 		if err != nil {
 			t.Fatalf("Could not make %s a KubernetesWorker: %v", nid, err)
@@ -163,12 +162,11 @@
 
 	// Remove KubernetesWorker from first node again. It will stay in k8s (arguably,
 	// this is a bug) but its role label should be removed.
-	no := false
 	_, err = mgmt.UpdateNodeRoles(ctx, &apb.UpdateNodeRolesRequest{
 		Node: &apb.UpdateNodeRolesRequest_Id{
 			Id: cluster.NodeIDs[0],
 		},
-		KubernetesWorker: &no,
+		KubernetesWorker: ptr.To(false),
 	})
 	if err != nil {
 		t.Fatalf("Could not remove KubernetesWorker from %s: %v", cluster.NodeIDs[0], err)