it's session mode and the pool size is 1200 (cause I need to grantee that in the worst case we have enough slots for all possible clients), however even at the times preceding high-cpu-sys-stall, the number postmasters are like 15-20. When stall happens, it starts to raise but that's the result of DB not answering in timely manner.
as far as data - yes, I can probably clean up names and provide some data with same queries, though - we are not able to cause a stall just with those queries & pgbench; we are able to make the situation worse and stall much likely when it's overlapped with regular queries coming from our app.