Understanding Status Metrics

A status Metric can evaluate a Monitor Test status result value and the apply the Metric evaluations Last N Times or Test Occurrences even when the Test does not return a Metric value.

A status Metric is added to the Test definition and automatically processed at the end of the normal Metric Test evaluation. The notification message is in the following format:

<Metric><Status Metric Name>=<test status>;Monitor=<Monitor Name>;Resource=<Resource Name>;Test=<Test Number></Metric>

<test status> has the following possible values:

• Severity1 = 9
• Severity2 = 8
• Severity3 = 7
• Severity4 = 6
• Bad = 5
• Warning = 4
• Good = 3

When evaluating the status Metric, the attributes for Monitor, Resource, and Test will be added and filtered, unless the evaluation always contains an attribute filter. An evaluation of 'max value = 3' will be automatically performed by default unless a Test Min, Test Max or Test Expression exists for the status Metric. Finally, the Last N Times or Test Occurrences evaluations will be performed if the Test status evaluation failed.

To enable the status Metric, set the Severity and select either Test Last N Times or Test Occurrences. You can use multiple evaluations to set different Severities.

A message is logged to the Monitor Log as shown in the Reveille Web interface for the case when the Monitor Test fails, but the status Metric evaluations succeed. This lets the user know that Tests are periodically failing, but the Tests are not marked as failed as they have not triggered the Last N Times or Test Occurrences evaluations. The message will appear similar to the following:

Status metric 'MonitorTest' evaluations succeeded for test failure at 2/21/2013 4:03:30 PM.