SAM AppInsight for IIS keeping worker processes alive

Question

Hello,

I tried to search through the forum archives to see if this topic had already been posted, but came up empty.

We have a handful of webservers that each host between dozens and hundreds of web apps each assigned to their own app pool. This design was inherited and having to maintain things this way for now. Most of the web apps get very little traffic so they are set to idle out after 20 minutes. However, once AppInsight for IIS is applied to a node, all the websites stay alive thanks to HTTP requests being sent by the monitor presumably as part of polling, so the worker processes never idle out anymore. As a result, the servers are now tying up significantly more resources as the all the worker processes are now always alive.

Basically, my question is whether the AppInsight of IIS template can be passive with its monitoring so the worker processes will timeout when not in use? For comparison, SCOM was already monitoring a couple of these webservers and is passive, yet will alert if an App Pool is down or unresponsive.

Any feedback is appreciated.

Scott Cummings

mesverrum · Accepted Answer

You might be able to stop the monitoring from keeping the workers up by disabling the http/s tests for the site.  All of the IIS metrics are coming back from powershell and WMI calls so it's just the actual page up/down connection tests that are keeping your workers up.