Alarms in Rhino alert the SLEE administrator to exceptional conditions.
Application components in the SLEE raise them, as does Rhino itself (upon detecting an error condition). Rhino clears some alarms automatically when the error conditions are resolved. The SLEE administrator must clear others manually.
When an alarm is raised or cleared, Rhino generates a JMX notification from the
Management clients may attach a notification listener to the Alarm MBean, to receive alarm notifications.
Rhino logs all alarm notifications.
While only SBBs could generate alarms in SLEE 1.0, other types of application components can also generate alarms in SLEE 1.1.
In SLEE 1.1, alarms are stateful — between being raised and cleared, an alarm persists in the SLEE, where an administrator may examine it. (In SLEE 1.0, alarms could be generated with a severity level that indicated a cleared alarm, but the fact that an error condition had occurred did not persist in the SLEE beyond the initial alarm generation.)
|For both SLEE 1.1 and 1.0, if the cause of an alarm is an Java exception, the log includes the exception and its stack trace (following the alarm description message).
|See also the How Do I Use Stateful Alarms in SLEE 1.1 on the OpenCloud Developer Portal.