Collecting user related metrics using JMX MBeans

There a few different MBeans to assist with tracking user related metrics that are useful for measuring adoption trends/growth as well usage that can be a factor in system load/performance. Consider adding these to available dashboards created to monitor your ELM applications. Complete details on their object names, attributes, collections, etc. can be found in the appropriate release's reference documentation.

Contributor Information MBean

This MBean includes attributes for

• totalCount - the total registered users in the repository
• unarchivedCount - how many of those users are active, that is not archived
• archivedCount - how may have been archived

This can help determine the extent to which your user population has been given access to a repository. If access is given gradually, that is, when needed, then the growth trend in this number can help measure adoption rate. Some clients find it easier to add all or large portions of their users en masse, though not necessarily assign them a license until needed; that method would make the measurement less helpful in showing adoption rate.

Comparing active user counts to license usage can provide good ratio metrics to help determine if you have sufficient licenses for future growth.

Active Services Summary MBean

This MBean includes an attribute for

• concurrentUsers - represents the number of unique users associated with an application's active services running at the time the bean data is collected

Concurrent users is an important data point to collect over time for growth planning/projection, to ensure you stay within known user scale maximums and to give insight into the load on an application when triaging an performance anomaly.

Server Activity Summary Metrics MBean

This MBean includes an attribute for

• activeUsersCount - represents the number of unique users associated with all of an application's active services that were running some time during the bean's collection interval (not the same as concurrent users)
• userWithMostServicesInInterval - of those active users, which one was the most active, in terms of total running services, during the collection interval

It is useful to track activeUsersCount to get a sense of what the 'normal' range of activity is for an application. When usage varies significantly beyond the normal range, look at other usage metrics, eg. CPU/memory utilization, to gauge whether the server is still running well within its constraints.

Monitoring userWithMostServicesInInterval may provide insight as to what user may be contributing to a performance anomaly or high load.