The CMIC collects operational events, problem events, and changes in system and component status for all components in the collectives it manages, and stores them in a database.
An event indicates a change in condition that may require intervention. The change may be good, bad, or insignificant to the ongoing operation of the system. For example, events include changes due to operations performed on the system, such as power off, and problem events reported to node event logs by node software.
- Teradata Database
- BYNET
- SCSI adapters
- Disk arrays
- Node operating systems
- Platform hardware
- Tape libraries
Summary alerts are groups of alerts known to occur together during a problem condition. The summary identifies the root cause, the components reporting the issue, and the recommended action. Alerts with a status of degraded or worse, which do not align with a known root cause, are grouped as Uncategorized Alerts.
Data bundles are packages of data used to diagnose problems. Data bundles are generated by software subsystems when they detect fault conditions. Data bundles reported during known fault conditions are automatically escalated to the TVI backend as soon as they are available, and become available as attachments to an associated summary alert.