The limitations of F5's LTM availability/resilience is defined only by how much you let LTM see. If LTM is position in front of the first tier (web server) then there is difficulty in what it can determine of application availability for the tiers behind that given level. Why? Because application availability is directly related to application visibility.
Consider a market trader, his produce is delivered daily. He can only sell what he sees at that immediate moment. However, a market trader that meets with the produce growers has insight into delivery today and in the future. And insight into trend delivers opportunity to react ahead of crisis.
This is no different for Application Delivery. F5's core development is to ensure fast and reliable delivery but even we struggle to make assurances for what we cannot see.
Ideally, Virtualise each tier of the App with an F5 LTM that is monitoring and measuring in real time. Alternatively, there will always be guess work.
Reference: http://www.f5.com/f5world/articles/volume8/F5World08-8-ways-to-virtualize.pdf
