How to solve "TCP retransmit timeout" & "TCP RST from remote system" issue on BIG-IP LTM?
Hi Experts,
We have an application which sends 80K+ https requests in 2 ~ 3 minutes. Client application extracts data from an API hosted behind BIG-IP . Big-IP is hosted in AWS EC2. Job was working fine for last couple of months but started failing for last three weeks. Usually, job runs for 40 minutes and extract all the data from API. No change was made on client application and on API. I captured packet in Big-IP and found that Big-IP is closing the connection with client, reasons are F5 TCP retransmit timeout and TCP RST from remote system. I applied one-connect profile with default parameter on the virtual server, now connection is not refused by Big-IP but job takes 20+ hours to extract the data from API. Applicaiton team increased backend AWS ECS instances (pool members) but no improvement. Please advise how to increase the connection on server side using one-connect profile. I haven't used it before.
Following is the topology:
Client APP ----> PaloAlto Firewall ---->(Internet)----> Big-IP(AWS EC2)----(Internet)---> AWS API gateway---> AWS NLB -----> AWS ECS.