diff --git a/clouddriver-aws/src/main/groovy/com/netflix/spinnaker/clouddriver/aws/provider/view/AmazonClusterProvider.groovy b/clouddriver-aws/src/main/groovy/com/netflix/spinnaker/clouddriver/aws/provider/view/AmazonClusterProvider.groovy
index 5dcfae3c348..46643ba6590 100644
--- a/clouddriver-aws/src/main/groovy/com/netflix/spinnaker/clouddriver/aws/provider/view/AmazonClusterProvider.groovy
+++ b/clouddriver-aws/src/main/groovy/com/netflix/spinnaker/clouddriver/aws/provider/view/AmazonClusterProvider.groovy
@@ -209,6 +209,14 @@ class AmazonClusterProvider implements ClusterProvider<AmazonCluster> {
         }
         instances.get(it)
       } ?: []
+
+      if (!serverGroup.instances && serverGroupEntry.attributes.instances) {
+        // has no direct instance relationships but we can partially populate instances based on attributes.instances
+        serverGroup.instances = serverGroupEntry.attributes.instances.collect {
+          new AmazonInstance(((Map) it) + [name: it.instanceId])
+        }
+      }
+
       [(serverGroupEntry.id) : serverGroup]
     }
 
diff --git a/clouddriver-core/src/main/groovy/com/netflix/spinnaker/clouddriver/model/ClusterProvider.groovy b/clouddriver-core/src/main/groovy/com/netflix/spinnaker/clouddriver/model/ClusterProvider.groovy
index e1053c0bdf4..810698199ad 100644
--- a/clouddriver-core/src/main/groovy/com/netflix/spinnaker/clouddriver/model/ClusterProvider.groovy
+++ b/clouddriver-core/src/main/groovy/com/netflix/spinnaker/clouddriver/model/ClusterProvider.groovy
@@ -100,6 +100,12 @@ interface ClusterProvider<T extends Cluster> {
    * Determines whether or not optimizations can be made by retrieving minimal or unexpanded clusters.
    *
    * This primarily affects how server groups are loaded for a cluster (@see com.netflix.spinnaker.clouddriver.controllers.ClusterController}.
+   *
+   * Minimal cluster support requires that server groups contained within a cluster be populated with:
+   * - creation time stamps
+   * - region / zone details
+   * - disabled status
+   * - instance counts (fully populated instances are not necessary)
    */
   boolean supportsMinimalClusters()
 }
diff --git a/clouddriver-web/src/main/groovy/com/netflix/spinnaker/clouddriver/controllers/ClusterController.groovy b/clouddriver-web/src/main/groovy/com/netflix/spinnaker/clouddriver/controllers/ClusterController.groovy
index f2b3ccce904..0f02e43402a 100644
--- a/clouddriver-web/src/main/groovy/com/netflix/spinnaker/clouddriver/controllers/ClusterController.groovy
+++ b/clouddriver-web/src/main/groovy/com/netflix/spinnaker/clouddriver/controllers/ClusterController.groovy
@@ -234,10 +234,6 @@ class ClusterController {
       )
     }
 
-    def expandServerGroups = { List<ServerGroup> serverGroups ->
-      return sortedServerGroups.collect { expandServerGroup(it) }
-    }
-
     switch (tsg) {
       case TargetServerGroup.CURRENT:
         return expandServerGroup(sortedServerGroups.get(0))
@@ -254,10 +250,10 @@ class ClusterController {
         return expandServerGroup(sortedServerGroups.last())
       case TargetServerGroup.LARGEST:
         // Choose the server group with the most instances, falling back to newest in the case of a tie.
-        return expandServerGroups(sortedServerGroups).sort { lhs, rhs ->
+        return expandServerGroup(sortedServerGroups.sort { lhs, rhs ->
           rhs.instances.size() <=> lhs.instances.size() ?:
-              rhs.createdTime <=> lhs.createdTime
-        }.get(0)
+            rhs.createdTime <=> lhs.createdTime
+        }.get(0))
       case TargetServerGroup.FAIL:
         if (sortedServerGroups.size() > 1) {
           throw new NotFoundException("More than one target found (scope: ${scope}, serverGroups: ${sortedServerGroups*.name})")