In our production system, queries sometimes 'stall'. Whilst stalled, no incrementing resource use (CPU, Reads) is shown in sp_whoisactive, and there is no blocking.
In retrospective diagnosis, we can see that sys.dm_db_stats_properties shows last_updated around the time the query was 'stalled'.
What we would like to do is - when we see a stalled query - then determine what auto stats updates are in progress.
Because we want to do this ad-hoc, and also because we don't want to impact production performance, using profiler is probably not an option for us.
(If there is no way of doing an ad-hoc determination, then maybe we'll have to consider Extended Events, or some other lower-impact pre-emptive tracking).
Our version is 2014, however answers for later versions would also be useful.