It might be the easier option—at least it will narrow down what the issue is. If it’s LiteFS related, we could try continuous profiling via profefe (although I haven’t personally used it before). Or, since it happens fairly regularly, you could sample the heap
profile from your home computer/laptop every 30 seconds:
while true
do
curl -s http://$HOSTNAME:20202/debug/pprof/heap > heap.`date -Iseconds`
sleep 30
end
Are you still seeing memory spikes?