Use this topic to learn about the alerts when a service is down.
The incoming alerts remain on the message bus until the fault management service acknowledges it to remove it off the message bus.
Fault Manager publishes alert notifications on the message bus. The notification service acknowledges all the alert notifications. If the notification service crashes or reboots, the un-acknowledged notifications that are left on the message bus, are published to the registered subscribers after the notification service reboots.
XCO does not persist messages across MQ reboot, so all pending alerts are lost. You can query the fault service for the missed alert notifications using the sequence IDs.
Note
Depending on the state of the message location, fault management service might not receive notifications from the components. Therefore, it does not raise an alert.
There are minimal chance of RabbitMQ rebooting alone . RabbitMQ reboots are usually associated with system issues which impact other services.
XCO attempts to re-notify fault management service on reboot. You can observe more frequent updates on HA status, storage status, and LDAP connectivity.
Most in-flight messages are lost. However, XCO ensures that alerts are regenerated and published on system restart. This also applies to failovers.
XCO increments the sequence ID correctly for alerts even after the system reboot and ensures ordered delivery of notifications.
A sub-system is a component of XCO that is responsible for publishing a message that is eventually converted into an alert by the fault management service.
Note
XCO cannot raise alerts during a sub-system reboot.