0

we host 4 web servers behind a shared load balancer. We don't have control over this Citrix NetScaler load balancer. We submit tickets to make changes as needed.

Recently, one of the web servers started to have issues like slow backend response times. Troubleshooting takes longer, therefore we wanted to have logic so if a request is not served within a certain threshold (let's say 15 secs) consider the server as down and forward requests to other servers from that time.

We contacted the hosting company and asked them whether it is possible to do this on their end. They say, it is provided via SNMP but they do not support SNMP on Shared load balancers.

We run the websites on IIS 8.5 (Server OS 2012 R2). We would like to know if any other IIS way of implementing this health check mechanism on our end exists?

Ismet D.
  • 1
  • 1
  • "They say, it is provided via SNMP" is vague. You'd better learn more about that before upgrading your hosting plan. Usually the load balancer pings dedicated health check pages to determine if an upstream server is available or not. So if your slow pages are not the health check pages, then the load balancer cannot help much on implementing the logic. IIS does not have such health check either. Hang dump analysis is a common way to identify root causes for such. You can open a support case via http://support.microsoft.com to learn more. – Lex Li Sep 03 '20 at 15:05
  • @LexLi are you referring to https://www.poppastring.com/blog/basic-hang-dump-analysis-using-windbg? – Ismet D. Sep 07 '20 at 18:28
  • That post says too little about the actual technique, but might help beginners to get some basic ideas. – Lex Li Sep 07 '20 at 19:17

0 Answers0