diff --git a/test_runner/performance/test_bulk_insert.py b/test_runner/performance/test_bulk_insert.py
index aafc315576..62301f3919 100644
--- a/test_runner/performance/test_bulk_insert.py
+++ b/test_runner/performance/test_bulk_insert.py
@@ -1,7 +1,7 @@
-import shutil
 from contextlib import closing
 
 from fixtures.compare_fixtures import NeonCompare, PgCompare
+from fixtures.pageserver.utils import wait_tenant_status_404
 from fixtures.pg_version import PgVersion
 
 
@@ -42,17 +42,14 @@ def measure_recovery_time(env: NeonCompare):
     client = env.env.pageserver.http_client()
     pg_version = PgVersion(client.timeline_detail(env.tenant, env.timeline)["pg_version"])
 
-    # Stop pageserver and remove tenant data
-    env.env.pageserver.stop()
-    timeline_dir = env.env.pageserver.timeline_dir(env.tenant, env.timeline)
-    shutil.rmtree(timeline_dir)
-
-    # Start pageserver
-    env.env.pageserver.start()
+    # Delete the Tenant in the pageserver: this will drop local and remote layers, such that
+    # when we "create" the Tenant again, we will replay the WAL from the beginning.
+    client.tenant_delete(env.tenant)
+    wait_tenant_status_404(client, env.tenant, iterations=60, interval=0.5)
+    client.tenant_create(new_tenant_id=env.tenant)
 
     # Measure recovery time
     with env.record_duration("wal_recovery"):
-        # Create the tenant, which will start walingest
         client.timeline_create(pg_version, env.tenant, env.timeline)
 
         # Flush, which will also wait for lsn to catch up