Scenario 4: Determining if the WebSphere cluster needs to be load balanced
Annette, the level 2 operator, gets a number of tickets relating to slow response time for an application. Annette receives an email indicating that the WASHighCPUPercentUsed situation triggered on the WebSphere® Application Server where the application is hosted.
- Annette navigates to the Tivoli® Enterprise Portal and observes that on the WebSphere Agent Summary workspace the Application icon is displaying a warning symbol. The Resources icon is displaying the critical symbol. The Resource icon flyover is displaying high JVM CPU%.
- Annette double-clicks the Resources icon and sees that the OS icon and JVM icon are both displaying the critical symbol. The OS icon flyover is displaying System CPU (ms) as high. The JVM icon also displays the JVM CPU% as high.
- Annette double-clicks the Application icon. The Application Trend at L2/L3 workspace is displayed. The Selected Application Summary report displays the application name, average request response time, average request completion rate, and error rate. The average request response time is high. The Request Rate Trend chart displays the number of requests that are completed per second for the application. Again, this value is displaying as high.
- Before Annette escalates this problem, she must determine if the
problem is recent or if it has been occurring for some time. Annette
checks the trend by taking the following steps in the Request Rate
Trend chart:
- In the Application Trend at L2/L3 workspace, she selects the Specify time span for query icon. The Select the Time Span window is displayed.
- In the Custom Parameters section, she enters the required values in the Start Time and End Time fields. Then, she clicks OK.
- Annette observes that there was an increase in client requests a few days ago and that this value has remained high throughout the week. Further investigation reveals that a surge of new customers caused a large increase in new users on the system. As a result, the load on the system is high.
- Annette uses an external ticketing tool to forward the ticket with all details to Jim, the Middleware/Application Support SME.
- Jim immediately sees from what Annette has reported that the system is over-burdened because of a significant increase in new users, and that the number of servers that are available in the cluster must be increased. Jim forwards the ticket to Simon, the OS SME.
- Simon must determine if the Application is running on a static or a dynamic WebSphere cluster. If the application is running on a static cluster, he adds additional application servers. If the application is running on a dynamic cluster, he increases the number of servers allowed. These actions are outside the scope of ITCAM for Application Diagnosticsthe agents.