Vitrage retrieves information from diverse data-sources regarding alarms in the cloud, and stores and combines this information into a unified system model. An important property of an alarm is its severity. This severity can be used both in the Vitrage templates to trigger actions, and in the Horizon UI for color-coding purposes. Therefore, it is important that within the Vitrage data model, severity names are aggregated and normalized.
Since each data-source might represent severity differently, for each
data-source we can supply it’s own mapping to the aggregated severity supported
in Vitrage. This way we can know which severity is more important.
In addition we also normalize the severities for the horizon UI (called
vitrage_operational_severity
) in order for the UI to know what color to show in
Horizon.
This page explains how to handle this mapping for a given data-source.
The alarm severity configuration is handled via config files. The location of these files can be determined in /etc/vitrage/vitrage.conf. Under the [entity_graph] section, set:
Name | Description | Default Value |
---|---|---|
datasources_values_dir | Directory path from where to load the values configurations | /etc/vitrage/datasources_values/ |
The alarm severity configuration files configure how the severity of each alarm is aggregated and normalized. There are several guidelines for creating a config file:
Once the file is modified, you must restart vitrage-graph service to load the changes.
Default configurations for alarms severities will be installed with Vitrage for all the pre-packaged data-sources.
category: ALARM
values:
- aggregated values:
priority: <Alarm severity priority - an integer>
original values:
- name: <Original alarm severity name>
operational_value: <normalized alarm severity - from OperationalAlarmSeverity class>
- name: ... # can list several severities for one aggregation
- aggregated values:
priority: ... # can list several aggregated severities
...
...
The following file will map alarm severities.
For aggregated severity with priority 40 we have 2 severities and each one of them is mapped to operational severity CRITICAL.
For aggregated severity with priority 30 we have 1 severity called WARNING and it is mapped to operational severity WARNING, etc…
category: ALARM
values:
- aggregated values:
priority: 40
original values:
- name: CRITICAL
operational_value: CRITICAL
- name: DOWN
operational_value: CRITICAL
- aggregated values:
priority: 30
original values:
- name: WARNING
operational_value: WARNING
- aggregated values:
priority: 20
original values:
- name: UNKNOWN
operational_value: N/A
- aggregated values:
priority: 10
original values:
- name: OK
operational_value: OK
- name: UP
operational_value: OK
Except where otherwise noted, this document is licensed under Creative Commons Attribution 3.0 License. See all OpenStack Legal Documents.