I found that the flow_id in the alert log may change after a while even though I keep sending the same packets (which means the packets are with the same src_port, src_addr, dst_port, dst_addr, proto).
I know that there’s a flow-timeouts configuration but I don’t think it’s the cause since I kept sending the packet without brreak.
So I think there may be some logic inside suricata that causes this behavior?
(note the id field in the screenshot is the flow_id field in eve log)