diff --git a/test/e2e/e2e.go b/test/e2e/e2e.go
index 1b27d8b7e1b..9b826b06112 100644
--- a/test/e2e/e2e.go
+++ b/test/e2e/e2e.go
@@ -184,7 +184,13 @@ func getDefaultClusterIPFamily(c clientset.Interface) string {
 // waitForDaemonSets for all daemonsets in the given namespace to be ready
 // (defined as all but 'allowedNotReadyNodes' pods associated with that
 // daemonset are ready).
+//
+// If allowedNotReadyNodes is -1, this method returns immediately without waiting.
 func waitForDaemonSets(c clientset.Interface, ns string, allowedNotReadyNodes int32, timeout time.Duration) error {
+	if allowedNotReadyNodes == -1 {
+		return nil
+	}
+
 	start := time.Now()
 	framework.Logf("Waiting up to %v for all daemonsets in namespace '%s' to start",
 		timeout, ns)
diff --git a/test/e2e/framework/node/wait.go b/test/e2e/framework/node/wait.go
index bc52a85b18d..be1c8ca0eb1 100644
--- a/test/e2e/framework/node/wait.go
+++ b/test/e2e/framework/node/wait.go
@@ -202,6 +202,9 @@ func checkWaitListSchedulableNodes(c clientset.Interface) (*v1.NodeList, error)
 func CheckReadyForTests(c clientset.Interface, nonblockingTaints string, allowedNotReadyNodes, largeClusterThreshold int) func() (bool, error) {
 	attempt := 0
 	return func() (bool, error) {
+		if allowedNotReadyNodes == -1 {
+			return true, nil
+		}
 		attempt++
 		var nodesNotReadyYet []v1.Node
 		opts := metav1.ListOptions{
diff --git a/test/e2e/framework/pod/wait.go b/test/e2e/framework/pod/wait.go
index 95e617697cd..8124540ba0e 100644
--- a/test/e2e/framework/pod/wait.go
+++ b/test/e2e/framework/pod/wait.go
@@ -102,7 +102,14 @@ func errorBadPodsStates(badPods []v1.Pod, desiredPods int, ns, desiredState stri
 // waiting. All pods that are in SUCCESS state are not counted.
 //
 // If ignoreLabels is not empty, pods matching this selector are ignored.
+//
+// If minPods or allowedNotReadyPods are -1, this method returns immediately
+// without waiting.
 func WaitForPodsRunningReady(c clientset.Interface, ns string, minPods, allowedNotReadyPods int32, timeout time.Duration, ignoreLabels map[string]string) error {
+	if minPods == -1 || allowedNotReadyPods == -1 {
+		return nil
+	}
+
 	ignoreSelector := labels.SelectorFromSet(map[string]string{})
 	start := time.Now()
 	e2elog.Logf("Waiting up to %v for all pods (need at least %d) in namespace '%s' to be running and ready",
diff --git a/test/e2e/framework/test_context.go b/test/e2e/framework/test_context.go
index b23942da406..679fa66e73b 100644
--- a/test/e2e/framework/test_context.go
+++ b/test/e2e/framework/test_context.go
@@ -289,7 +289,7 @@ func RegisterCommonFlags(flags *flag.FlagSet) {
 	flags.StringVar(&TestContext.LogexporterGCSPath, "logexporter-gcs-path", "", "Path to the GCS artifacts directory to dump logs from nodes. Logexporter gets enabled if this is non-empty.")
 	flags.BoolVar(&TestContext.DeleteNamespace, "delete-namespace", true, "If true tests will delete namespace after completion. It is only designed to make debugging easier, DO NOT turn it off by default.")
 	flags.BoolVar(&TestContext.DeleteNamespaceOnFailure, "delete-namespace-on-failure", true, "If true, framework will delete test namespace on failure. Used only during test debugging.")
-	flags.IntVar(&TestContext.AllowedNotReadyNodes, "allowed-not-ready-nodes", 0, "If non-zero, framework will allow for that many non-ready nodes when checking for all ready nodes.")
+	flags.IntVar(&TestContext.AllowedNotReadyNodes, "allowed-not-ready-nodes", 0, "If greater than zero, framework will allow for that many non-ready nodes when checking for all ready nodes. If -1, no waiting will be performed for ready nodes or daemonset pods.")
 
 	flags.StringVar(&TestContext.Host, "host", "", fmt.Sprintf("The host, or apiserver, to connect to. Will default to %s if this argument and --kubeconfig are not set.", defaultHost))
 	flags.StringVar(&TestContext.ReportPrefix, "report-prefix", "", "Optional prefix for JUnit XML reports. Default is empty, which doesn't prepend anything to the default name.")
@@ -352,7 +352,7 @@ func RegisterClusterFlags(flags *flag.FlagSet) {
 
 	flags.StringVar(&cloudConfig.ClusterTag, "cluster-tag", "", "Tag used to identify resources.  Only required if provider is aws.")
 	flags.StringVar(&cloudConfig.ConfigFile, "cloud-config-file", "", "Cloud config file.  Only required if provider is azure or vsphere.")
-	flags.IntVar(&TestContext.MinStartupPods, "minStartupPods", 0, "The number of pods which we need to see in 'Running' state with a 'Ready' condition of true, before we try running tests. This is useful in any cluster which needs some base pod-based services running before it can be used.")
+	flags.IntVar(&TestContext.MinStartupPods, "minStartupPods", 0, "The number of pods which we need to see in 'Running' state with a 'Ready' condition of true, before we try running tests. This is useful in any cluster which needs some base pod-based services running before it can be used. If set to -1, no pods are checked and tests run straight away.")
 	flags.DurationVar(&TestContext.SystemPodsStartupTimeout, "system-pods-startup-timeout", 10*time.Minute, "Timeout for waiting for all system pods to be running before starting tests.")
 	flags.DurationVar(&TestContext.NodeSchedulableTimeout, "node-schedulable-timeout", 30*time.Minute, "Timeout for waiting for all nodes to be schedulable.")
 	flags.DurationVar(&TestContext.SystemDaemonsetStartupTimeout, "system-daemonsets-startup-timeout", 5*time.Minute, "Timeout for waiting for all system daemonsets to be ready.")
diff --git a/test/e2e/framework/util.go b/test/e2e/framework/util.go
index a239aa0586e..2040604a816 100644
--- a/test/e2e/framework/util.go
+++ b/test/e2e/framework/util.go
@@ -1024,8 +1024,11 @@ func getNodeEvents(c clientset.Interface, nodeName string) []v1.Event {
 // WaitForAllNodesSchedulable waits up to timeout for all
 // (but TestContext.AllowedNotReadyNodes) to become schedulable.
 func WaitForAllNodesSchedulable(c clientset.Interface, timeout time.Duration) error {
-	Logf("Waiting up to %v for all (but %d) nodes to be schedulable", timeout, TestContext.AllowedNotReadyNodes)
+	if TestContext.AllowedNotReadyNodes == -1 {
+		return nil
+	}
 
+	Logf("Waiting up to %v for all (but %d) nodes to be schedulable", timeout, TestContext.AllowedNotReadyNodes)
 	return wait.PollImmediate(
 		30*time.Second,
 		timeout,
@@ -1118,11 +1121,16 @@ func RunHostCmdWithRetries(ns, name, cmd string, interval, timeout time.Duration
 	}
 }
 
-// AllNodesReady checks whether all registered nodes are ready.
+// AllNodesReady checks whether all registered nodes are ready. Setting -1 on
+// TestContext.AllowedNotReadyNodes will bypass the post test node readiness check.
 // TODO: we should change the AllNodesReady call in AfterEach to WaitForAllNodesHealthy,
 // and figure out how to do it in a configurable way, as we can't expect all setups to run
 // default test add-ons.
 func AllNodesReady(c clientset.Interface, timeout time.Duration) error {
+	if TestContext.AllowedNotReadyNodes == -1 {
+		return nil
+	}
+
 	Logf("Waiting up to %v for all (but %d) nodes to be ready", timeout, TestContext.AllowedNotReadyNodes)
 
 	var notReady []*v1.Node