This was a tool that I developed within my own time to support activities targeting some odd behaviour within our Production eMatrix environments.
It’s basically a dashboard view of the production floor, combining various tools to make the following information easily available: –
- Application Server
- The number of Long Running Threads vs total threads on the application server
- The used memory vs total available memory on the application servers
- The average throughput the application is sending across the network (since being started)
- The last reboot time of the Application Servers
- The number of Long Running Threads vs total threads in Matrix
- The used memory vs total available memory on the matrix RMIs
- Ability to terminating long running matrix threads through the interface.
- The number of Long Running Threads vs total threads in Oracle
- The number of disk partitions on the machine which are in a (configurable per partition) “Warning” or “Error” state, based on % used / total space
- Provides Interface based alerts (pop-up dialogues from the system tray upon an event)
This application required no modification on the servers themselves – it runs totally stand-alone on your local desktop allowing the key support personnel to keeps track of the system health.
The tool also supported exporting the state of the full environment to XML. One document detailed all the Oracle session data and thread dumps from the 12 active Java virtual machines. To couple this export I also developed an XQuery database that allowed us to run trend analysis and get to the root cause of the issues we were experiencing.
Overall it was a fun little home project that provided us with a very helpful tool in resolving the issues we were experiencing.