System Monitoring

Monitoring Internal OIM Processes

Yale's Big Brother system has been configured to monitor OIM by checking that the JBoss process is running on the server. This, however, does not tell us if the OIM system itself is OK. Therefore, a status page has been added to OIM that provides for checking the following:

• Connection pools:  We've found that these can occasionally run out of connections if there is a downstream problem.  OIM does not always recover these connections when the underlying problem has been addressed.
• Provisioning Task Errors:  Short of logging into the Admin Console, OIM provides no direct means of telling if there are provisioning errors or not.  Additionally, all provisioning errors are treated as an error, even though a simple retry may resolve the problem.  Yale has developed a scheme of classifying provisioning errors so that BB may alert us when critical issues arrise that need our attention (such as duplicate SSNs), or when an error can be ignored (such as the system being Quiesced).
• Scheduled Tasks:  Should a recon fail, error out, or fail to run as scheduled, our process will notify BB of the problem so we can address the situation.

The system provides a Detail page that lists all the monitored items and their status.  There is also a Overall page that omits all the detail.  If any monitored item reports a Trouble status, the overall status is Trouble.  Otherwise, if any monitored item reports an Attention status, the overall status is attention.  Otherwise, the overall status is OK.

The home monitoring page also provides convenient links to related websites:

• Magic Tool
• OIM Admin Console
• The IDM Wiki
• The IDM Jira
• Tomcat Status
• JMX Console
• JBoss Web Console
• The Monitoring Page for Patch, Dev, Test and Production
• The current developer contacts

Documents