eG Help

Using the WebLogic server's SNMP interface, this test tracks the rate of requests processed by the server, the number of requests waiting for processing, and the percentage of heap usage by the server. While the rate of requests processed and the number of queued requests can be indicative of performance problems with the WebLogic server, the percentage heap usage can be indicative of the reason for the problem.

The heap size determines how often, and for how long garbage collection is performed by the Java Virtual Machine (JVM) that hosts the WebLogic server. The Java heap is a repository for live objects, dead objects, and free memory. When the JVM runs out of memory in the heap, all execution in the JVM stops while a Garbage Collection (GC) algorithm goes through memory and frees space that is no longer required by an application. This is an obvious performance hit because users accessing a WebLogic server must wait while GC happens. No server-side work can be done during GC. Consequently, the heap size must be tuned to minimize the amount of time that the JVM spends in garbage collection, while at the same time maximizing the number of clients that the server can handle at a given time. For Java 2 environments, it is recommended that the heap size be set to be as possible without causing the host system to "swap" pages to disk (use the output of eG's SystemTest to gauge the amount of swapping being performed by the operating system). The measures made by this test are as follows:

Measurement	Description	Measurement Unit	Interpretation
Throughput	Rate of requests processed by the WebLogic server.	Reqs/Sec	A high request rate is an indicator of server overload. By comparing the request rates across application servers, an operator can gauge the effectiveness of load balancers (if any) that are in use.
Pct_heap_used	Percentage of heap space currently in use by the WebLogic server.	Percent	When the heap used percent reaches 100%, the server will start garbage collection rather than processing requests. Hence, a very high percentage of heap usage (close to 100%) will dramatically lower performance. In such a case, consider increasing the heap size to be used.
Queued_requests	Number of requests currently waiting to be processed by the server.	Number	An increase in number of queued requests can indicate a bottleneck on the WebLogic server. One of the reasons for this could be a bottleneck at the server or due one or more of the applications hosted on the server.
Total_heap_size	Current heap size of the WebLogic server's Java Virtual Machine	MB
Free_heap_size	The currently unused portion of the WebLogic server's Java Virtual Machine	MB