22-Jun-2023 13:10
I changed Idle Timeout form 300 secs to a bigger value mainly followed below steps , but when traffic is following through F5 , user is seeing timeout in around 5 mins when they check directly from Pool member it is more than 5 mins.
What wrong I am doing ? How can I validate this ?
https://my.f5.com/manage/s/article/K7166
Creating a new protocol profile that uses a different timeout value
Note: For BIG-IP 11.5.0 and later versions, clicking Advanced is no longer necessary.
The virtual server now uses the new idle timeout setting.
22-Jun-2023 13:28
You may want to review the following article:
https://my.f5.com/manage/s/article/K13004262
Also, note that many times firewalls will also have a default 300s timeout. So also make sure any other network device in the path has been amended in terms of its timeout config.
22-Jun-2023 14:42
My Keep Alive Interval - is higher than Idle Timeout , since idle time out should trigger before
22-Jun-2023 15:02
Hi @Subrun if you take a verbose tcpdump capture w/ the F5 flags enabled, you'll be able to see the reset reason on the packets in the capture, if it is the big-ip that is resetting it and not an upstream devices that has shorter timers.
24-Jun-2023 11:14
My understanding is that keep alive should be shorter. So if the client in indeed still connected, a response to keep alive extends the amount of time remaining until idle timeout.
If an application is enterprise grade and developed by a 3rd party, they usually have guidance for integration with app delivery controllers like F5. They should be able to provide guidance on what values to set on the Virtual Server config for the app.