Node Monitoring vs Pool Monitoring

As you have discovered, you can monitor application health as delivered through a BIG-IP LTM system at either the node or pool member level, with all sorts of different strategies and tactics available. Therefore, the short answer to your question about how monitoring works is, "It depends." For example, if a monitor that is assigned to a node marks the node as down, then all pool members that reference that node will also be marked down, having inherited that status from their parent node. If all members in a pool are marked down, the pool's status also becomes down. Likewise, if that is the only pool a particular virtual server has associated with it, the virtual server may also be marked down.

When it comes down to it though, your approach to application health monitoring will depend on a lot of factors, not the least of which is the ability of your network to absorb the additional traffic generated by monitoring while still maintaining good throughput and performance for your clients. This is a pretty big topic, but also a well documented one. F5 offers both free and paid training that covers application health monitoring with BIG-IP LTM. And, yes, there are also quite a few DevCentral and AskF5 articles that cover monitoring. Here are a few links for both training and documentation resources:

• Getting Started with BIG-IP (Part 2): Application Delivery (free)
• Digital Education Series: LTM: Monitoring Application Delivery Health (self-paced; paid)
• Administering BIG-IP (2 days, paid)
• Configuring BIG-IP LTM (3 days, paid)
• BIG-IP Local Traffic Manager: Monitors Reference
• BIG-IP System: Maintaining High Availability through Resource Monitoring