CA SystemEDGE User Guide › Concepts › State Management Model › Object Aggregation

Object Aggregation

The object class, instance, and attribute values in the Self Monitor and Process Monitor tables describe the monitored (or managed) object. Multiple monitors within same object instance, class, and attribute values are considered to be connected and describe the same managed object. The agent aggregates the worst state of all connected monitors into a single managed object state. The agent creates an entry (aggregate) for each aggregation in the Aggregate table.

However, if two connected monitors A and B have the same severity and use “AND” flag (00800 in monFlags or 20000 in pmonFlags), A and B are excluded from worst state calculation, if A and B do not have the same state. If you do not enter values into the corresponding columns, the agent populates them with meaningful default information. It bases default self monitor values on the monitored OID and maps them to class, instance, and attribute values. It bases default process monitor values on the process regular expression and monitored attribute. For more information about default object information assignment, see the chapters "Self Monitoring" and "Process and Service Monitoring."

All connected monitors must have the same interval which helps ensure that the aggregate state is evaluated only once for all connected monitors and prevents unnecessary state changes.

The Aggregate table contains the following information:

• Object class, instance, and attribute values
• Row status
• Values for the monitor entry with the worst aggregated state, such as threshold, operator, and current value
• Current aggregate state
• Monitor statistics, such as the number of state change traps sent, the time spent in any of the possible states, and the previous status

Note: For more information about the columns in the Aggregate table, see the chapter "Self Monitoring."

The agent updates the current state in the Aggregate table. The agent sends a state change trap only when the threshold evaluation of all connected monitors and the subsequent state aggregation leads to a different worst state than the previous one. For example, assume the agent returns 82% CPU usage, while you monitor CPU usage with three monitors: one for 60% (assigned a warning severity), one for 80% (critical severity), and one for 100% (fatal severity). This value causes a threshold breach for the 60% and 80% monitors, but the agent only sends one state change trap for the 80% monitor and changes the aggregate state to critical.