Earlier this week I spent some time helping a company troubleshoot some performance issues with Windows Azure – their ASP.NET request queue was growing longer than the maximum amount of queued requests supported by default in IIS (5000) during bursts of high activity, so I had to help them find a way to empty the queue faster.

One way to do this is to modify the number of max/min IO and worker threads available to any web role instance, and these values can only be modified through machine.config settings.

As you know, Windows Azure VMs are not persistent so you can’t modify machine.config and save it to file system like how you might if you were hosting IIS on your own local box.

Thus, I crafted a solution using AppCmd and Windows Azure startup tasks.

Here’s some of the code below:

REM Increases the number of available IIS threads for high performance applications REM Uses the recommended values from http://msdn.microsoft.com/en-us/library/ms998549.aspx#scalenetchapt06_topic8 REM Values may be subject to change depending on your needs REM Assumes you're running on two cores (medium instance on Azure) %windir%\system32\inetsrv\appcmd set config /commit:MACHINE -section:processModel -maxWorkerThreads:100 >> log.txt 2>> err.txt %windir%\system32\inetsrv\appcmd set config /commit:MACHINE -section:processModel -minWorkerThreads:50 >> log.txt 2>> err.txt %windir%\system32\inetsrv\appcmd set config /commit:MACHINE -section:processModel -minIoThreads:50 >> log.txt 2>> err.txt %windir%\system32\inetsrv\appcmd set config /commit:MACHINE -section:processModel -maxIoThreads:100 >> log.txt 2>> err.txt

I’ve open-sourced the rest of my IIS machine.config / app pool scripts on Github. Take a look at them and let me know if you have any questions.

If you enjoyed this post, make sure you subscribe to my RSS feed!