slapos_cloud: fix script_ComputeNode_reportBang unauthorized error

No need to access invalidated instances.

Also try to reduce conflict, and try to not block indexation activities.

master/bt5/slapos_cloud/WorkflowTemplateItem/portal_workflow/compute_node_slap_interface_workflow/script_ComputeNode_reportBang.py

 compute_node = state_change['object']
 portal = compute_node.getPortalObject()
-for compute_partition in [x for x in compute_node.contentValues(portal_type='Compute Partition') if x.getSlapState() == 'busy']:
-  for instance_sql in portal.portal_catalog(
-                        default_aggregate_uid=compute_partition.getUid(),
-                        portal_type=["Software Instance", "Slave Instance"],
-                        ):
-    instance = instance_sql.getObject()
-    if instance.getSlapState() in ["start_requested", "stop_requested"]:
-      instance.activate().SoftwareInstance_bangAsSelf(
-        relative_url=instance.getRelativeUrl(),
-        reference=instance.getReference(), 
-        comment=state_change.kwargs.get('comment', ''))
+compute_partition_uid_list = [x.getUid() for x in compute_node.contentValues(portal_type='Compute Partition') if x.getSlapState() == 'busy']
+
+for instance_sql in portal.portal_catalog(
+  default_aggregate_uid=compute_partition_uid_list,
+  portal_type=["Software Instance", "Slave Instance"],
+  validation_state='validated',
+  # Try limiting conflicts on multiple instances of the same tree
+  # example: monitor frontend
+  group_by_list=['specialise_uid']
+):
+  instance = instance_sql.getObject()
+  if instance.getSlapState() in ["start_requested", "stop_requested"]:
+    # Increase priority to not block indexations
+    # if there are many activities created
+    instance.activate(priority=2).SoftwareInstance_bangAsSelf(
+      relative_url=instance.getRelativeUrl(),
+      reference=instance.getReference(),
+      comment=state_change.kwargs.get('comment', '')
+    )
 
 compute_node.setErrorStatus('bang')