diff --git a/cluster-autoscaler/core/static_autoscaler.go b/cluster-autoscaler/core/static_autoscaler.go
index 3dfced108c82..3e537f14d764 100644
--- a/cluster-autoscaler/core/static_autoscaler.go
+++ b/cluster-autoscaler/core/static_autoscaler.go
@@ -246,6 +246,11 @@ func (a *StaticAutoscaler) RunOnce(currentTime time.Time) errors.AutoscalerError
 		return nil
 	}
 
+	// Update cluster resource usage metrics
+	coresTotal, memoryTotal := calculateCoresMemoryTotal(allNodes, currentTime)
+	metrics.UpdateClusterCoresCount(coresTotal)
+	metrics.UpdateClusterMemoryCount(memoryTotal)
+
 	daemonsets, err := a.ListerRegistry.DaemonSetLister().List(labels.Everything())
 	if err != nil {
 		klog.Errorf("Failed to get daemonset list: %v", err)
@@ -800,3 +805,21 @@ func getUpcomingNodeInfos(registry *clusterstate.ClusterStateRegistry, nodeInfos
 	}
 	return upcomingNodes
 }
+
+func calculateCoresMemoryTotal(nodes []*apiv1.Node, timestamp time.Time) (int64, int64) {
+	// this function is essentially similar to the calculateScaleDownCoresMemoryTotal
+	// we want to check all nodes, aside from those deleting, to sum the cluster resource usage.
+	var coresTotal, memoryTotal int64
+	for _, node := range nodes {
+		if isNodeBeingDeleted(node, timestamp) {
+			// Nodes being deleted do not count towards total cluster resources
+			continue
+		}
+		cores, memory := core_utils.GetNodeCoresAndMemory(node)
+
+		coresTotal += cores
+		memoryTotal += memory
+	}
+
+	return coresTotal, memoryTotal
+}
diff --git a/cluster-autoscaler/metrics/metrics.go b/cluster-autoscaler/metrics/metrics.go
index 782ba930e77e..ea1bc2bad601 100644
--- a/cluster-autoscaler/metrics/metrics.go
+++ b/cluster-autoscaler/metrics/metrics.go
@@ -138,6 +138,14 @@ var (
 		},
 	)
 
+	clusterCoresCount = k8smetrics.NewGauge(
+		&k8smetrics.GaugeOpts{
+			Namespace: caNamespace,
+			Name:      "cluster_cores_count",
+			Help:      "Current number of cores in the cluster, minus deleting nodes.",
+		},
+	)
+
 	coresLimitsCount = k8smetrics.NewGaugeVec(
 		&k8smetrics.GaugeOpts{
 			Namespace: caNamespace,
@@ -146,6 +154,14 @@ var (
 		}, []string{"direction"},
 	)
 
+	clusterMemoryCount = k8smetrics.NewGauge(
+		&k8smetrics.GaugeOpts{
+			Namespace: caNamespace,
+			Name:      "cluster_memory_count",
+			Help:      "Current number of bytes of memory in the cluster, minus deleting nodes.",
+		},
+	)
+
 	memoryLimitsCount = k8smetrics.NewGaugeVec(
 		&k8smetrics.GaugeOpts{
 			Namespace: caNamespace,
@@ -304,7 +320,9 @@ func RegisterAll() {
 	legacyregistry.MustRegister(nodeGroupsCount)
 	legacyregistry.MustRegister(unschedulablePodsCount)
 	legacyregistry.MustRegister(maxNodesCount)
+	legacyregistry.MustRegister(clusterCoresCount)
 	legacyregistry.MustRegister(coresLimitsCount)
+	legacyregistry.MustRegister(clusterMemoryCount)
 	legacyregistry.MustRegister(memoryLimitsCount)
 	legacyregistry.MustRegister(lastActivity)
 	legacyregistry.MustRegister(functionDuration)
@@ -382,12 +400,22 @@ func UpdateMaxNodesCount(nodesCount int) {
 	maxNodesCount.Set(float64(nodesCount))
 }
 
+// UpdateClusterCoresCount records the number of cores in the cluster, minus deleting nodes
+func UpdateClusterCoresCount(coresCount int64) {
+	clusterCoresCount.Set(float64(coresCount))
+}
+
 // UpdateCoresLimitsCount records the minimum and maximum number of cores in the cluster
 func UpdateCoresLimitsCount(minCoresCount int64, maxCoresCount int64) {
 	coresLimitsCount.WithLabelValues("minimum").Set(float64(minCoresCount))
 	coresLimitsCount.WithLabelValues("maximum").Set(float64(maxCoresCount))
 }
 
+// UpdateClusterMemoryCount records the number of bytes of memory in the cluster, minus deleting nodes
+func UpdateClusterMemoryCount(memoryCount int64) {
+	clusterMemoryCount.Set(float64(memoryCount))
+}
+
 // UpdateMemoryLimitsCount records the minimum and maximum bytes of memory in the cluster
 func UpdateMemoryLimitsCount(minMemoryCount int64, maxMemoryCount int64) {
 	memoryLimitsCount.WithLabelValues("minimum").Set(float64(minMemoryCount))
diff --git a/cluster-autoscaler/proposals/metrics.md b/cluster-autoscaler/proposals/metrics.md
index 3b535f46948e..31d7062ad5ee 100644
--- a/cluster-autoscaler/proposals/metrics.md
+++ b/cluster-autoscaler/proposals/metrics.md
@@ -28,7 +28,9 @@ All the metrics are prefixed with `cluster_autoscaler_`.
 | unschedulable_pods_count | Gauge | | Number of unschedulable ("Pending") pods in the cluster. |
 | node_groups_count | Gauge | `node_group_type`=<node-group-type> | Number of node groups managed by CA. |
 | max_nodes_count | Gauge | | Maximum number of nodes in all node groups. |
+| cluster_cores_count | Gauge | | | Current number of cores in the cluster, minus deleting nodes. |
 | cores_limits_count | Gauge | `direction`=<`minimum` or `maximum`> | Minimum and maximum number of cores in the cluster. |
+| cluster_memory_count | Gauge | | Current number of bytes of memory in the cluster, minus deleting nodes. |
 | memory_limits_count | Gauge | `direction`=<`minimum` or `maximum`> | Minimum and maximum number of bytes of memory in cluster. |
 
 * `cluster_safe_to_autoscale` indicates whether cluster is healthy enough for autoscaling. CA stops all operations if significant number of nodes are unready (by default 33% as of CA 0.5.4).