1

we've got some old legacy web apps that we host on an azure VM. In some circumstances they get stuck in a situation that spins the W3WP process into 99%> and causes grief for the server. They're not mission critical so it's safer for us to auto kill these processes when this occurs rather then opening up the code and changing old code.

We've tested the IIS Limit process functionality locally and it works perfectly, after 3 minutes of the process running > 90% the w3wp process is killed.

<cpu limit="90000" action="KillW3wp" resetInterval="00:03:00" />

I run the exact same settings on the VM in azure and it doesnt do anything. The process will continue to run high and IIS wont kick in and kick it out.

Any ideas?

Matt
  • 119
  • 3
  • Could you outline what the applications are doing? Connecting to DBs? – Papa May 23 '16 at 00:48
  • Hi Papa, Thanks for the reply. It's just an old .net wcf service connecting to a db. I wouldnt have thought this would have much to do with what was executing in the app. I was more thinking why this should be any different between our local IIS instances vs an IIS instance in the cloud on the Azure Classic VM? – Matt May 23 '16 at 00:55
  • TBH, some things just act differently in VMs. What are the specs of you VM? Could be IO related. – Papa May 24 '16 at 00:36
  • Is there anyway you can profile the service? See what it's doing? – Papa May 24 '16 at 00:40
  • @Matt did you ever get this figured out? – TWilly May 11 '17 at 18:23

0 Answers0