diff --git a/k8s/argo-workflows/deluge-benchmark-workflow.yaml b/k8s/argo-workflows/deluge-benchmark-workflow.yaml
index 6ec845d..6d4e5f8 100644
--- a/k8s/argo-workflows/deluge-benchmark-workflow.yaml
+++ b/k8s/argo-workflows/deluge-benchmark-workflow.yaml
@@ -27,7 +27,7 @@ spec:
         value: "false"
       # If set to false, leaves pods for failed experiments behind so they can be inspected.
       - name: cleanupOnFailure
-        value: "false"
+        value: "true"
       # If set to false, does not parse/upload logs at the end of the experiment.
       - name: parseLogs
         value: "true"