diff --git a/dashboards/cluster.jsonnet b/dashboards/cluster.jsonnet
index a6fb2ef..cca5b29 100755
--- a/dashboards/cluster.jsonnet
+++ b/dashboards/cluster.jsonnet
@@ -150,7 +150,7 @@ local nodepoolCPUCommitment =
       Most commonly, JupyterHub workloads are *memory bound*, not CPU bound. So this is
       not a particularly helpful graph.
 
-      Common shapes:
+      Common shapes this visualization may take:
       1. If this is *consistently high* but shaped differently than your memory commitment
         graph, consider changing your CPU requirements.
     |||