There's an option to have an IIS Application Pool automatically kill w3wp if the process runs at over some threshold for some period. Ref: Kill w3wp.exe if it is working on 100% cpu over a period of time
When such an event occurs, I'd like to trigger a mail such that we're aware that there had been an issue; thus we'll know:
- any users reporting issues before the alert will likely have had their issue resolved by the KillW3wp event.
- any users reporting issues at the point of KillW3wp would suggest unwanted side effects caused by the kill.
- we can search for patterns in when this fix is needed / in data in the related logs from the same time, from which we can delve deeper into finding the underlying cause.
- Is it possible to trigger an email when KillW3wp is initiated?
- Alternatively, does anyone know the EventId, Source & Log we'd need to monitor to detect this issue (we could thus use the windows scheduler to trigger the mail; http://blogs.technet.com/b/jhoward/archive/2010/06/16/getting-event-log-contents-by-email-on-an-event-log-trigger.aspx)