On one of my SharePoint servers (that has 24 CPU cores and 134 GB of RAM) the worker process for our main SharePoint application will occassionally reach 98% - 100% CPU utilization and bring the whole farm to a crawl. This generally starts happening during full search crawls.
Until I actually figure out what is causing the resource utilization, I wanted your thoughts on this temporary work around:
Creating a process that monitors the CPU utilization of the IIS worker processes. If one sustains a CPU utilization of 90% or more for more than a minute, recycle the application pool for that process automatically.
Does anyone see any issues with doing that?