Monitoring the Coyote Point Equalizer

eG Enterprise offers a specialized Coyote Point Equalizer (see Figure 1) monitoring model, which involves a single eG external agent that periodically polls the SNMP MIB of the equalizer, and collects a wide variety of performance information revealing the load on the device and the effectiveness with which the device balances this load across the servers in a farm. In the event of inconsistencies in load balancing, the agent proactively alerts administrators to the potential problem, so that he/she can initiate the relevant remedial action immediately.

Figure 1 : The layer model of the Coyote Point Equalizer

Each layer of Figure 1 above is mapped to tests that report the following:

• How many clusters are being managed by the equalizer and what are they? Is any cluster overloaded currently? If so, which one is it?
• Which cluster is currently handling the maximum number of connections?
• Which cluster is the busiest in terms of hits to its servers?
• How is the connection load on the equalizer? Is the equalizer able to handle the load?
• Which type of connections is the highest on the equalizer -  Level-4 or Level-7?
• Did any connection to the equalizer time out?
• Is the equalizer evenly distributing load across all the servers in the cluster, or is any server currently overloaded?
• Is the equalizer able to assure requests of quick responses from the servers, or is any server in the cluster responding slowly to client requests? Is it owing to a badly tuned equalizer?
• Are client connections to a cluster uniformly distributed across all the servers in that cluster? If not, what is the reason for the imbalance? 
• Is any server in the cluster idle?