eG Monitoring
 

Measures reported by CpsServerStatusTest

The real test of the efficiency of a load balancer lies in its ability to uniformly distribute load across the servers in a cluster, thereby ensuring the peak performance and continuous availability of the dependent services. Using the Equalizer Server Status test, administrators can accurately judge the efficiency and effectiveness of the equalizer. This test monitors the connection and calculated load on each server in a cluster, promptly detects load imbalances, and alerts administrators to them, so that they can quickly resolve the issue.

This test is used to monitors the connection and calculated load on each server in a cluster, promptly detects load imbalances, and alerts administrators to them, so that they can quickly resolve the issue

 The measures made by this test are as follows:

Measurement Description Measurement Unit Interpretation
Server_load Indicates the current calculated load value for this server. Number This indicates the workload on the server. By comparing the value of this measure across all the servers in a cluster, you can instantly identify irregularities in load balancing. If found necessary, you can reconfigure the load balancing rules to ensure uniform load distribution across servers.
Response_time Indicates how quickly this server is currently responding to client requests. ms It is the job of a load balancer to ensure minimal response time for client requests. A high value for this measure could therefore indicate a defective load balancer or one that is improperly configured. Further investigation is hence necessary in this case to identify the root-cause of the anomaly.
Current_connections Indicates the number of connections that were active on this server during the last measurement period. Number The indicates the connection load on the server. By observing the graph of this measure over time, you can analyze the rate of growth of the load on the server. By comparing the value of this measure across all the servers in a cluster, you can instantly identify overloaded servers; this in turn brings irregularities in load balancing to light.
Total_connections Indicates the number of current connections to this server. Number If a sudden/consistent increase in the value of this measure is noticed, you might have to investigate further to identify the reason for this occurrence.
Idle_time Indicates the time for which this server was idle. secs Ideally, the value of this measure should be low. A high value indicates that the server has remained unused for a long time. This could be owing to inconsistencies in load balancing or because the server is unavailable for use.