Pool crashed showing lots of http work queue depth exceeded

[Leonardiae]

2022-04-26T04:39:29Z WARNING: request rejected because http work queue depth exceeded, it can be increased with the -rpcworkqueue= setting
2022-04-26T04:39:29Z WARNING: request rejected because http work queue depth exceeded, it can be increased with the -rpcworkqueue= setting
2022-04-26T04:39:30Z WARNING: request rejected because http work queue depth exceeded, it can be increased with the -rpcworkqueue= setting
2022-04-26T04:39:30Z WARNING: request rejected because http work queue depth exceeded, it can be increased with the -rpcworkqueue= setting

I had configured the setting in the btcsq daemon used by the pool in btcsq.conf to:
rpcworkqueue=10000

[Warchant]

Don't make work queue too long. Maybe 100 or less is best value. It should not crash with small values.

[overcookedpanda]

I experienced this bug today with APM and pool connected to the node using a value of 256 which I have used in the past with most pools, will try to re-configure using 100, but it seems with lower values it still occurs.

[reb0rn21]

Don't make work queue too long. Maybe 100 or less is best value. It should not crash with small values.

if you put less then ~1000 when there is a lot POP on network, node just respond that RPC queue is at limit, I tried less and more, even pushed open files in linux to more as its crash the node with open file limit

for me when i see the btcsq process use one full core CPU 100% (and mostly it does) its just matter of them when node will get slow and RPC will get full queue
that need to be sorted out why 100% load it can not work as that, or it need to scale work to multicore as CPU is not issue (its ryzen 3600)

I am not a coder at all but BTC mem pool is limited to I think 300MB and if more txs come in it will look only at ones with more fees, I am not sure how they solve txs flood in mempool as POP has similar mempool issue is somewhere in there