diff --git a/test_runner/performance/pageserver/pagebench/test_pageserver_max_throughput_getpage_at_latest_lsn.py b/test_runner/performance/pageserver/pagebench/test_pageserver_max_throughput_getpage_at_latest_lsn.py
index 772a39fe35..68f3d9dcbe 100644
--- a/test_runner/performance/pageserver/pagebench/test_pageserver_max_throughput_getpage_at_latest_lsn.py
+++ b/test_runner/performance/pageserver/pagebench/test_pageserver_max_throughput_getpage_at_latest_lsn.py
@@ -209,3 +209,11 @@ def run_benchmark_max_throughput_latest_lsn(
             unit="ms",
             report=MetricReport.LOWER_IS_BETTER,
         )
+
+    env.storage_controller.allowed_errors.append(
+        # The test setup swaps NeonEnv instances, hence different
+        # pg instances are used for the storage controller db. This means
+        # the storage controller doesn't know about the nodes mentioned
+        # in attachments.json at start-up.
+        ".* Scheduler missing node 1",
+    )