Very interesting, indeed. We are currently checking for quorum in the iRule with active_members across all of the pools but as we increase user and ajax calls >5K we see a noticeable increase in LTM CPU from all of this checking. We are trying to decouple the monitoring from the user traffic so as not to increase monitoring when user traffic increases, and to have finite resources dedicated to monitoring.
The core problem for this post is the inelegant method to determine Pool member status via all of the command line tools, including tmsh. I am looking for the best method to determine how many members of a pool are available to take traffic -- ie UP and not disabled. Currently it seems that you have to read in the output of these commands and look in two places to make sure the node can actually take traffic. I was hoping that someone has figured out a better way...
-Dan