摘要 |
A method for diagnosing problems in a computer system by visualizing flows through applications and other subsystems in a directed graph on a user interface. The user interface represents multiple instances of each application or other subsystem by a respective node (502, 820, 401, 504, 821, 826, 828, 832, 834, 836), and edges (523, 524, 853) indicate which nodes (502, 820, 401, 504, 821, 826, 828, 832, 834, 836) depend on one another. Aggregate metrics which are based on the multiple instances, and associated alerts, can be provided for the nodes (502, 820, 401, 504, 821, 826, 828, 832, 834, 836) and edges (523, 524, 853). An aging process can indicate which nodes have not been recently invoked. The user interface can also indicate which nodes (502, 820, 401, 504, 821, 826, 828, 832, 834, 836) and edges (523, 524, 853) are associated with a given business transaction (BTA1, 804, 806, 808, 810, 812). In a summary view, a node hides the identity of invoked components such as servlets of the application, while in a detailed view these details are provided. |