Collecting basic health check metrics using JMX MBeans

The ELM application provides a basic health check widget you can add to a dashboard. In addition, there a few different MBeans you can monitor related to base application health.

Server Information MBean

The bean provides some health information for the server running the application. Of most interest are:

• dbConnectionStatus - indicates if the status of the application's connection to the database
• dbPingTime - a measure of the latency between the application and database

A working and low latency connection to the database are essential to the application providing a good quality of service. Monitor these attributes to detect if the connection fails or the latency degrades appreciably over time.

Repotools Verify Information MBean

This bean provides the results of an online verify that checks the integrity of an application's production database. There are several verifiers that execute, for each, the attributes to watch are

• componentId - identifies the validator being run
• statusCode - indicates the result of the particular verifier
• severity - provides an assessment of the severity of verification results, eg. OK, WARNING, ERROR and others.

This is a resource intensive scenario so should be run infrequently and during off hours. Monitor the results to detect any reported errors to further investigate, with IBM Support, as needed.

Diagnostics MBean

Server diagnostics can be run regularly to analyze many aspects of the server so as to quickly identify problems. There are several tests that execute, for each, the attributes to watch are

• testId - the test that ran
• status - the overall status of the test execution
• statusDesc - short description of the test results
• detailStatusDesc - more information about the test results

Monitor the results to detect any reported errors to further investigate, with IBM Support, as needed. Use of this bean is also described in CLM Monitoring.

LogEvents MBean

This bean tracks error and warning messages captured in the application log file. Key errors to watch for are documented in Monitoring error messages. Occurrence of these messages should be investigated, with IBM Support, as needed.

Some useful attributes include

• errorId - the ID of the error
• errorTitle - title of the error
• errorDetails - additional details for the error
• userId - the user involved in the request that resulted in the error

Some clients have log parsing enabled in their application monitoring environment. These likely provide a more robust capability than what this MBean provides.