Fix flaky test_volume_metrics

ref: 7626

Signed-off-by: Chris <[email protected]>

manager/integration/tests/test_metric.py

@@ -7,7 +7,7 @@
 
 from common import client, core_api, volume_name  # NOQA
 
-from common import crash_replica_processes
+from common import delete_replica_processes
 from common import create_pv_for_volume
 from common import create_pvc_for_volume
 from common import create_snapshot
@@ -287,7 +287,7 @@ def test_volume_metrics(client, core_api, volume_name, pvc_namespace): # NOQA
 
     volume.updateReplicaCount(replicaCount=3)
     volume = wait_for_volume_healthy(client, volume_name)
-    crash_replica_processes(client, core_api, volume_name)
+    delete_replica_processes(client, core_api, volume_name)
     volume = wait_for_volume_faulted(client, volume_name)
 
     check_metric(core_api, "longhorn_volume_robustness",