Er - please disregard my last comment about ‘similar callstacks’ - as it turns out it looks like the second call is happening due to some other as yet unknown (to me) async path. It may or may not be a the issue but hopefully will provide a hint / breadcrumb to whomever is digging into it.