diff --git a/node/nodeutil/controller.go b/node/nodeutil/controller.go
index f3b3205dd..96d1a8aa5 100644
--- a/node/nodeutil/controller.go
+++ b/node/nodeutil/controller.go
@@ -254,6 +254,11 @@ type NodeConfig struct {
 	// Set the error handler for node status update failures
 	NodeStatusUpdateErrorHandler node.ErrorHandler
 
+	// SkipDownwardAPIResolution can be used to skip any attempts at resolving downward API references
+	// in pods before calling CreatePod on the provider.
+	// Providers need this if they need to do their own custom resolving
+	SkipDownwardAPIResolution bool
+
 	routeAttacher func(Provider, NodeConfig, corev1listers.PodLister)
 }
 
@@ -393,13 +398,14 @@ func NewNode(name string, newProvider NewProviderFunc, opts ...NodeOpt) (*Node,
 	}
 
 	pc, err := node.NewPodController(node.PodControllerConfig{
-		PodClient:         cfg.Client.CoreV1(),
-		EventRecorder:     cfg.EventRecorder,
-		Provider:          p,
-		PodInformer:       podInformer,
-		SecretInformer:    secretInformer,
-		ConfigMapInformer: configMapInformer,
-		ServiceInformer:   serviceInformer,
+		PodClient:                 cfg.Client.CoreV1(),
+		EventRecorder:             cfg.EventRecorder,
+		Provider:                  p,
+		PodInformer:               podInformer,
+		SecretInformer:            secretInformer,
+		ConfigMapInformer:         configMapInformer,
+		ServiceInformer:           serviceInformer,
+		SkipDownwardAPIResolution: cfg.SkipDownwardAPIResolution,
 	})
 	if err != nil {
 		return nil, errors.Wrap(err, "error creating pod controller")
diff --git a/node/pod.go b/node/pod.go
index b0b0b9ad6..ecee718b0 100644
--- a/node/pod.go
+++ b/node/pod.go
@@ -69,11 +69,13 @@ func (pc *PodController) createOrUpdatePod(ctx context.Context, pod *corev1.Pod)
 		"namespace": pod.GetNamespace(),
 	})
 
-	// We do this so we don't mutate the pod from the informer cache
-	pod = pod.DeepCopy()
-	if err := podutils.PopulateEnvironmentVariables(ctx, pod, pc.resourceManager, pc.recorder); err != nil {
-		span.SetStatus(err)
-		return err
+	if !pc.skipDownwardAPIResolution {
+		// We do this so we don't mutate the pod from the informer cache
+		pod = pod.DeepCopy()
+		if err := podutils.PopulateEnvironmentVariables(ctx, pod, pc.resourceManager, pc.recorder); err != nil {
+			span.SetStatus(err)
+			return err
+		}
 	}
 
 	// We have to use a  different pod that we pass to the provider than the one that gets used in handleProviderError
diff --git a/node/podcontroller.go b/node/podcontroller.go
index c31a05ea3..97c263a04 100644
--- a/node/podcontroller.go
+++ b/node/podcontroller.go
@@ -140,6 +140,11 @@ type PodController struct {
 	// This is used since `pc.Run()` is typically called in a goroutine and managing
 	// this can be non-trivial for callers.
 	err error
+
+	// skipDownwardAPIResolution can be used to skip any attempts at resolving downward API references
+	// in pods before calling CreatePod on the provider.
+	// Providers need this if they need to do their own custom resolving
+	skipDownwardAPIResolution bool
 }
 
 type knownPod struct {
@@ -196,6 +201,11 @@ type PodControllerConfig struct {
 	// For example, if the pod informer is not filtering based on pod.Spec.NodeName, you should
 	// set that filter here so the pod controller does not handle events for pods assigned to other nodes.
 	PodEventFilterFunc PodEventFilterFunc
+
+	// SkipDownwardAPIResolution can be used to skip any attempts at resolving downward API references
+	// in pods before calling CreatePod on the provider.
+	// Providers need this if they need to do their own custom resolving
+	SkipDownwardAPIResolution bool
 }
 
 // NewPodController creates a new pod controller with the provided config.
@@ -236,15 +246,16 @@ func NewPodController(cfg PodControllerConfig) (*PodController, error) {
 	}
 
 	pc := &PodController{
-		client:             cfg.PodClient,
-		podsInformer:       cfg.PodInformer,
-		podsLister:         cfg.PodInformer.Lister(),
-		provider:           cfg.Provider,
-		resourceManager:    rm,
-		ready:              make(chan struct{}),
-		done:               make(chan struct{}),
-		recorder:           cfg.EventRecorder,
-		podEventFilterFunc: cfg.PodEventFilterFunc,
+		client:                    cfg.PodClient,
+		podsInformer:              cfg.PodInformer,
+		podsLister:                cfg.PodInformer.Lister(),
+		provider:                  cfg.Provider,
+		resourceManager:           rm,
+		ready:                     make(chan struct{}),
+		done:                      make(chan struct{}),
+		recorder:                  cfg.EventRecorder,
+		podEventFilterFunc:        cfg.PodEventFilterFunc,
+		skipDownwardAPIResolution: cfg.SkipDownwardAPIResolution,
 	}
 
 	pc.syncPodsFromKubernetes = queue.New(cfg.SyncPodsFromKubernetesRateLimiter, "syncPodsFromKubernetes", pc.syncPodFromKubernetesHandler, cfg.SyncPodsFromKubernetesShouldRetryFunc)