Alarms can be raised from any part of the system and be dispatched, in real-time, to any target (e.g. e-mail, SNMP trigger, etc.) through the Alarm Manager. The criteria for conditions that should be treated as alarms (as opposed to normal errors) are completely configurable by the operator.
By using an Event Notification configuration it will be possible to filter and stream events of all types and levels produced by various MediationZone functions in real-time, as data is processed through the system. Valid destinations are files on the local disk, MediationZone system log, and e-mail notifications. The following events are managed:
Code Server events
Node failures
Node messages
Node state changes
System startup, shutdown, and failures
Task Executor messages
Task state changes
User actions
Workflow state changes
Events can be generated in a format recognized by fault management and performance management systems, e.g. an SNMP trap.
MediationZone provides for tracing and debug capabilities within its standard functionality. Descriptions and recommendations of “what-to-do” are provided as part of the standard documentation.