eG Help

A WebLogic server (prior to version 9.x) may be configured with different execute queues. By default, a WebLogic server is configured with one thread queue that is used for execution by all applications running on a server instance. A common way of improving a WebLogic server's performance is by configuring multiple thread execute queues. For eg., a mission-critical application can be assigned to a specific thread execute queue, thereby guaranteeing it a fixed number of execute threads. Other, less critical applications may compete for threads in the default execute queue. While using different thread execute queues can significantly improve performance, if the thread execute queues are not properly configured or maintained, this could result in less than optimal performance. For eg., you may find that while one thread queue has a number of idle threads, applications in another thread execute queue could be waiting for execute threads to become available. In case of the WebLogic server prior to version 9.x, the WebLogicThreads test monitors the different thread execute queues configured for the server.

From WebLogic server 9.x onwards however, execute queues are replaced by 'work managers'. Therefore, while monitoring WebLogic server 9.x or above, the WebLogicThreads test will report one set of metrics for every 'work manager' configured for the server. Also, the test will take an additional ThreadPool descriptor, which will report the extent of usage of the thread pool.

Measurement	Description	Measurement Unit	Interpretation
Idle_threads	Indicates the number of idle threads in the thread pool	Number	This measure will be available only for the ThreadPool descriptor. If the value of this measure nears the maximum number of threads in the WebLogic server, it is an indication of excessive activity in the server. It also indicates a probable delay in the processing of subsequent requests.
Pct_thread_utilization	Indicates the percentage of threads utilized in a pool	Percent	>This measure will be available only for the ThreadPool descriptor. When this value becomes 100 %, it indicates a heavy load on the server and that it cannot process further requests until a few threads become idle.
Pending_requests	Indicates the number of requests waiting in the queue / work manager	Number	A high value of this measure is indicative of a processing delay.
Request_rate	Indicates the number of requests that are processed by the queue / work manager per second	Reqs/sec	While a high value of this measure is indicative of the good health of the server, a low value indicates a processing bottle-neck.