Monitor Pool member CPU in Load Balancing scheme

If you use something other than round robin for the default lb method, particularly Observed or Predictive, and an application health check that is better than just a ping or tcp port check, then most nodes won't get into this situation in the first place. The LTM will be able to respond dynamically to the load on the nodes as evidenced by their application response time, and you won't have to load agents on the nodes to poll CPU and such.

 

 

Denny