diff --git a/pkg/kubelet/container/runtime.go b/pkg/kubelet/container/runtime.go
index c7ea7029cda..5618aba2298 100644
--- a/pkg/kubelet/container/runtime.go
+++ b/pkg/kubelet/container/runtime.go
@@ -81,6 +81,16 @@ type Runtime interface {
 	GetContainerLogs(containerID, tail string, follow bool, stdout, stderr io.Writer) (err error)
 }
 
+// Customizable hooks injected into container runtimes.
+type RuntimeHooks interface {
+	// Determines whether the runtime should pull the specified container's image.
+	ShouldPullImage(pod *api.Pod, container *api.Container, imagePresent bool) bool
+
+	// Runs after an image is pulled reporting its status. Error may be nil
+	// for a successful pull.
+	ReportImagePull(pod *api.Pod, container *api.Container, err error)
+}
+
 // Pod is a group of containers, with the status of the pod.
 type Pod struct {
 	// The ID of the pod, which can be used to retrieve a particular pod
diff --git a/pkg/kubelet/kubelet.go b/pkg/kubelet/kubelet.go
index e26ae61aa7e..a0d2ca07536 100644
--- a/pkg/kubelet/kubelet.go
+++ b/pkg/kubelet/kubelet.go
@@ -235,6 +235,7 @@ func NewMainKubelet(
 		resourceContainer:              resourceContainer,
 		os:                             osInterface,
 		oomWatcher:                     oomWatcher,
+		runtimeHooks:                   newKubeletRuntimeHooks(recorder),
 	}
 
 	if plug, err := network.InitNetworkPlugin(networkPlugins, networkPluginName, &networkHost{klet}); err != nil {
@@ -402,8 +403,14 @@ type Kubelet struct {
 	// Name must be absolute.
 	resourceContainer string
 
-	os         kubecontainer.OSInterface
+	os kubecontainer.OSInterface
+
+	// Watcher of out of memory events.
 	oomWatcher OOMWatcher
+
+	// TODO(vmarmol): Remove this when we only have to inject the hooks into the runtimes.
+	// Hooks injected into the container runtime.
+	runtimeHooks kubecontainer.RuntimeHooks
 }
 
 // getRootDir returns the full path to the directory under which kubelet can
@@ -867,41 +874,25 @@ func parseResolvConf(reader io.Reader) (nameservers []string, searches []string,
 
 // Pull the image for the specified pod and container.
 func (kl *Kubelet) pullImage(pod *api.Pod, container *api.Container) error {
-	if container.ImagePullPolicy == api.PullNever {
-		return nil
-	}
-
-	start := time.Now()
-	defer func() {
-		metrics.ImagePullLatency.Observe(metrics.SinceInMicroseconds(start))
-	}()
-
-	ref, err := kubecontainer.GenerateContainerRef(pod, container)
-	if err != nil {
-		glog.Errorf("Couldn't make a ref to pod %v, container %v: '%v'", pod.Name, container.Name, err)
-	}
 	present, err := kl.containerManager.IsImagePresent(container.Image)
 	if err != nil {
+		ref, err := kubecontainer.GenerateContainerRef(pod, container)
+		if err != nil {
+			glog.Errorf("Couldn't make a ref to pod %v, container %v: '%v'", pod.Name, container.Name, err)
+		}
 		if ref != nil {
 			kl.recorder.Eventf(ref, "failed", "Failed to inspect image %q: %v", container.Image, err)
 		}
 		return fmt.Errorf("failed to inspect image %q: %v", container.Image, err)
 	}
 
-	if container.ImagePullPolicy == api.PullAlways ||
-		(container.ImagePullPolicy == api.PullIfNotPresent && (!present)) {
-		if err := kl.containerManager.Pull(container.Image); err != nil {
-			if ref != nil {
-				kl.recorder.Eventf(ref, "failed", "Failed to pull image %q: %v", container.Image, err)
-			}
-			return err
-		}
-		if ref != nil {
-			kl.recorder.Eventf(ref, "pulled", "Successfully pulled image %q", container.Image)
-		}
+	if !kl.runtimeHooks.ShouldPullImage(pod, container, present) {
+		return nil
 	}
 
-	return nil
+	err = kl.containerManager.Pull(container.Image)
+	kl.runtimeHooks.ReportImagePull(pod, container, err)
+	return err
 }
 
 // Kill all running containers in a pod (includes the pod infra container).
diff --git a/pkg/kubelet/kubelet_test.go b/pkg/kubelet/kubelet_test.go
index 2b42f8da1cf..804bc25ad14 100644
--- a/pkg/kubelet/kubelet_test.go
+++ b/pkg/kubelet/kubelet_test.go
@@ -123,6 +123,7 @@ func newTestKubelet(t *testing.T) *TestKubelet {
 	kubelet.containerManager.Prober = kubelet.prober
 	kubelet.handlerRunner = newHandlerRunner(&fakeHTTP{}, &fakeContainerCommandRunner{}, kubelet.containerManager)
 	kubelet.volumeManager = newVolumeManager()
+	kubelet.runtimeHooks = newKubeletRuntimeHooks(kubelet.recorder)
 	return &TestKubelet{kubelet, fakeDocker, mockCadvisor, fakeKubeClient, waitGroup, fakeMirrorClient}
 }
 
diff --git a/pkg/kubelet/metrics/metrics.go b/pkg/kubelet/metrics/metrics.go
index 6452642fd2c..8ee6083da3f 100644
--- a/pkg/kubelet/metrics/metrics.go
+++ b/pkg/kubelet/metrics/metrics.go
@@ -28,13 +28,6 @@ import (
 const kubeletSubsystem = "kubelet"
 
 var (
-	ImagePullLatency = prometheus.NewSummary(
-		prometheus.SummaryOpts{
-			Subsystem: kubeletSubsystem,
-			Name:      "image_pull_latency_microseconds",
-			Help:      "Image pull latency in microseconds.",
-		},
-	)
 	ContainersPerPodCount = prometheus.NewSummary(
 		prometheus.SummaryOpts{
 			Subsystem: kubeletSubsystem,
@@ -73,7 +66,6 @@ var registerMetrics sync.Once
 func Register(containerCache kubecontainer.RuntimeCache) {
 	// Register the metrics.
 	registerMetrics.Do(func() {
-		prometheus.MustRegister(ImagePullLatency)
 		prometheus.MustRegister(SyncPodLatency)
 		prometheus.MustRegister(DockerOperationsLatency)
 		prometheus.MustRegister(SyncPodsLatency)
diff --git a/pkg/kubelet/runonce_test.go b/pkg/kubelet/runonce_test.go
index 6bcb39bb80d..c0d1084a380 100644
--- a/pkg/kubelet/runonce_test.go
+++ b/pkg/kubelet/runonce_test.go
@@ -90,6 +90,7 @@ func TestRunOnce(t *testing.T) {
 		os:                  kubecontainer.FakeOS{},
 		volumeManager:       newVolumeManager(),
 	}
+	kb.runtimeHooks = newKubeletRuntimeHooks(kb.recorder)
 
 	kb.networkPlugin, _ = network.InitNetworkPlugin([]network.NetworkPlugin{}, "", network.NewFakeHost(nil))
 	if err := kb.setupDataDirs(); err != nil {
diff --git a/pkg/kubelet/runtime_hooks.go b/pkg/kubelet/runtime_hooks.go
new file mode 100644
index 00000000000..1b2a70a7a38
--- /dev/null
+++ b/pkg/kubelet/runtime_hooks.go
@@ -0,0 +1,64 @@
+/*
+Copyright 2015 Google Inc. All rights reserved.
+
+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 kubelet
+
+import (
+	"github.com/GoogleCloudPlatform/kubernetes/pkg/api"
+	"github.com/GoogleCloudPlatform/kubernetes/pkg/client/record"
+	kubecontainer "github.com/GoogleCloudPlatform/kubernetes/pkg/kubelet/container"
+	"github.com/golang/glog"
+)
+
+// Kubelet-specific runtime hooks.
+type kubeletRuntimeHooks struct {
+	recorder record.EventRecorder
+}
+
+var _ kubecontainer.RuntimeHooks = &kubeletRuntimeHooks{}
+
+func newKubeletRuntimeHooks(recorder record.EventRecorder) kubecontainer.RuntimeHooks {
+	return &kubeletRuntimeHooks{
+		recorder: recorder,
+	}
+}
+
+func (kr *kubeletRuntimeHooks) ShouldPullImage(pod *api.Pod, container *api.Container, imagePresent bool) bool {
+	if container.ImagePullPolicy == api.PullNever {
+		return false
+	}
+
+	if container.ImagePullPolicy == api.PullAlways ||
+		(container.ImagePullPolicy == api.PullIfNotPresent && (!imagePresent)) {
+		return true
+	}
+
+	return false
+}
+
+func (kr *kubeletRuntimeHooks) ReportImagePull(pod *api.Pod, container *api.Container, pullError error) {
+	ref, err := kubecontainer.GenerateContainerRef(pod, container)
+	if err != nil {
+		glog.Errorf("Couldn't make a ref to pod %q, container %q: '%v'", pod.Name, container.Name, err)
+		return
+	}
+
+	if pullError != nil {
+		kr.recorder.Eventf(ref, "failed", "Failed to pull image %q: %v", container.Image, pullError)
+	} else {
+		kr.recorder.Eventf(ref, "pulled", "Successfully pulled image %q", container.Image)
+	}
+}