Monitoring F5 BIG-IP Load Balancers

The eG Enterprise provides a specialized F5 BigIp model (see Figure 1) for managing a BIG-IP load balancer.

Figure 1 : The layer model of a BIG-IP load balancer

Every layer of the layer model depicted by Figure 1 is mapped to one/more tests, which uses the host's SNMP MIB to extract critical statistics pertaining to the load balancing activity performed by BIG-IP. The performance metrics reported by these measures reveal the following:

System Monitoring
  • Is the Big-IP hardware sufficiently sized to handle the incoming load, or is there a CPU or a memory bottleneck?
  • Is I/O activity on the system abnormal?
  • Are context switches kept at a minimum, or are there too many threads contending for CPU resources?
  • How much swap space is currently available in the system? Is it sufficient?
Workload Monitoring
  • What is the current total workload on the Big-IP load balancer? Has there been any change in the workload over time?
  • How much traffic is the load balancer handling? How many connections
Load Distribution monitoring
  • Has any of the servers across which the load is being balanced failed?
  • Is network load balanced across all the servers in the pool?
  • Is any server in the pool handling more connection requests than others?