Configuring load balancing down to app. pool level


Our organization is using the A10 Thunder 3030S to load balance two Windows VM Servers (on the same VM host) that are hosting an IIS Web App. The load balancer is using a round robin algorithm with sticky sessions enabled (Source IP Persistence) between the client and server.

Using the round robin algorithm, the load balancer only detects if a server is down but is unaware if the application pool is down.

1. Is it possible to enable load balancing down to the application pool level either by creating a health check or is there a recommended/common setup to follow?
2. Since we are only load balancing IIS application pools, should we be using an HTTP protocol or is TCP fine? Our web. App uses Single Sign-on and when we change the protocol to HTTP, SSO stops working.

I've attached screenshots of our current setup.

Any information or advice anyone who has a similar setup would be greatly appreciated.

Apologies in advance, I am a novice user working with our in-house support to understand what options are available and how others have configured their load balancers to achieve similar goal/setup.



  • hkohn72hkohn72 Member
    edited March 2018
    Hello Jason,

    thank you for your detailed description.

    1. It is possible and recommended to create a health monitor on service group level. You find it in the deployment guide
    2. At TCP loadbalancing you only do the loadbalancing at network layer 4 and can not look into the application. With HTTP loadbalancing you do it on layer 7 and have full view into the content. It is better to check and work on layer 7 because you can use WAF and more aFlex funtions on the traffic.

    If you need further information or help for your discussion please contact us(A10 Networks).

    CU hkohn!
Sign In or Register to comment.