This custom dashboard provides a filterable, sortable table of App Server JVM metrics with drilldowns.
Full size: http://dl.dropbox.com/u/3077237/JVMHeap104.avi 9mb
Use the filter in the top left of the screen to specify a subset of all Java Systems to display.
The filter in the top right filters by app server or host name, or, using the "Advanced" button, by metric value.
App servers that match the specified filters are shown in the table with columns representing the alarm state of the app server, name, host name, current heap%, current MB used, period avg MB used, max heap capacity, GC rate (period), GC overhead (Period), request alarm health, current request count, and current request response time. Each column is sortable and has a configured dwell view and selection popup. You can click on elements of the selection popup to drill down to an appropriate diagnostic view.
The out-of-the-box Java dashboards do a great job of navigating from the top level to any particular app server, application, resource or request, but there is no easy way to compare resources from multiple app servers without going to the page associated with the first app server, noting the value, then going to the second app server, then comparing the value.
This dashboard easily answers questions like:
- which app server is in the worst state?
- which app server has the highest heap utilization?
- which app server has the highest garbage collection overhead?
- which app server has the slowest average response time?
- which app server has the highest load?
- are my load balanced app servers all processing the same number of requests, or are they unbalanced?
The cartridge attached is ready to install on any FMS version 5.5.4 and higher.
Watch the video for details.
Use (Please visit the site to view this file) for Java cartridge 5.8.0+.
Use (Please visit the site to view this file) for FMS 5.5.8 with Java Cartridge 5.7.4 and lower..
Use (Please visit the site to view this file) for FMS 5.5.4.
Update JVMHeap 5.8.1.1 so that request metrics for Tomcat are visible on the dashboard.
Known issues:
- Oracle App Servers do not appear in the dashboard - workaround described in comments below.
- Response times for Tomcat are shown in milliseconds rather than seconds.
Other dashboards by Brian Wheeldon
Custom Dashboards for Device Availability metrics captured by the NetMonitor agent
Common Dashboard Framework for Custom Script Agents and Utility OS Agents
Custom Terminal Services Dashboard
Foglight 5.6 Infrastructure Host property dashboards and (re)portlets
Building a custom dashboard to summarize LogFilter alarms
Customizing the out-of-the-box Host Status view
(Please visit the site to view this file)
(Please visit the site to view this file)
(Please visit the site to view this file)
(Please visit the site to view this file)