When we tested VS application, we found that occasionally the connection was rejected, and it would appear 1-2 times a day. But this problem does not occur when the client directly accesses the server. The F5 is configured with the FastL4 VS. There are two pool members in the Pool. Others are regular configurations. After the tm.rstcause.log switch is turned on on f5, when the test phenomenon is reappeared, the RST log is not found on f5. How can I troubleshoot this issue?
Is this a long lived connection? What could be happening is the BIG-IP resetting the connection due to innactivity. Have you tried bumping the idle-timeout up?