We have faced sudden memory spikes in our StarRocks BE nodes twice in the last couple of days. Due to this our pods end up restarting (we have deployed our cluster using Kubernetes operator). Is there a way to figure out which table, which row or which process is causing the spike? Error message is attached

No direct way to see it. You can see the general mem usage via http

curl "<http://127.0.0.1:8040/mem_tracker>"

or inside use this query

show data;

Too many tablet usually lead to higher memory consumption