we do some stress test on one of our VIP (no more than 10 TPS) and our client get from time to time RST connection , we also see correlation between this reset to the HTTP monitor frequency although the node/pool doesn’t marked as down.in trace we did on F5 we see ACK and after 4 second RST but we never see the retransmission
have you checked rst cause?
sol13223: Configuring the BIG-IP system to log TCP RST packets
Please check if the reset is caused on one or all specific pool member's.
Check with netstat command on pool member's
and also confirm if the timeout setting over server\node and F5 virtual server are same..
sometimes if they are different, there might be possibility for increasing TCP RST packet over server or F5 end
Have you tried using a different monitor or the same monitor with a different interval and checked the correlation?