| 摘要 |
If an event management system or component within an event management system fails, events generated by systems and applications monitored by the event management system may not be processed or analyzed. As a result, the performance and health of the systems and applications may be unknown. To reduce the impact of a failure, an event management system can include at least two event managers: a primary event manager and one or more standby event managers. The standby event managers are configured to handle event processing tasks upon a failure of the primary event manager. To be prepared for a failure of the primary event manager, the standby event manager receives state data and copies of events. After determining that the primary event manager has failed, the standby event manager can begin processing events. |
| 主权项 |
1. A method comprising:
receiving, by both a first event manager and a second event manager, a first batch of events from a communication bus, wherein the communication bus received events, that include the first batch of events, from a plurality of components of a network; communicating, by the first event manager to the second event manager, first state data that at least indicates state of processing of the first batch of events by the first event manager; after communication of the first state data from the first event manager to the second event manager,
the first event manager discarding the first batch of events;without processing the first batch of events, the second event manager discarding the first batch of events; andreceiving, by both the first event manager and the second event manager, a second batch of events from the communication bus; determining, by the second event manager, that the first event manager is not operational; and in response to determining that the first event manager is not operational,
processing, by the second event manager, the second batch of events based, at least in part, on the first state data; andupdating, by the second event manager, the first state data based, at least in part, on processing the second batch of events. |
