Scenario: 1 web server out of 4 demonstrates cpu at 100%, where all others max out at 90%

I recently cloned a new web server (2008r2, iis6) and added it into a production workload (load balanced with NetScaler)

During a recent situation where a SaS web application demonstrated slow response times, I became aware of a capacity issue - the root cause of the application slow down

Our monitoring solution - zabbix - indicated 3 webservers (the original 3) maxed out their cpu at 90% whereas the fourth, new web server hit 100%

This behaviour suggested a setting had not been included in the clone process that was allowing the fourth webserver to use all of its CPU where the other 3 had been set to only allow hit 90% and leave 10% free

I have been unable to find the setting that configures this and can't think of any other places to look after checking the following:

IIS - Application Pool - Advanced settings (Limit + Limit Action etc) Windows Resource Manager IIS Config file - Application.Host XenServer (the virtualization host)

Can anyone suggest other places to look or have any experience in this kind of scenario?


You can't natively throttle the cpu per process.

You need to check what use the cpu and ajust according to that.

On the other hands please check the cpu configuration for the server versus other server. In example I seen monitoring tool display 50% cpu usage (for a server with 2vcpu) while in fact the cpu hit 100% for the process but for only one core. So it was a monitoring tool missconfig, as when the process hit 100%, it stop answering request, even if another cpu/core is ready. It could explain the result in Zabbix.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.