SLA reports detail how effectively hosts met their Service Level Agreement targets.

Given a set of hosts (specified by a set of hostnames or hostname wild card, or group names or group wildcard), a datasource is selected, instances selected (wildcard matches allowed), and a specific metric is selected to evaluate the SLA against.

The report then requires the user to specify what constitutes an acceptable service level.

e.g. to generate an SLA report showing the amount of time the selected hosts had at most 80% CPU load:

 

 

Note: the resultant report shows the amount of time the hosts passed the SLA - not the amount of CPU load.

Note: the granularity of the SLA report is currently limited by the data aggregation functions (i.e to 5 minute periods within the last week, and 3 hour periods within the last month.) This means that any SLA violation may be under reported if the average metric during the aggregation period passed the SLA. For example, an SLA report showing a failure for more than 10 % ping loss (PingLossPercent < 10) may show no failures for a host, even though it had a 20% ping loss for 9 minutes, three weeks ago. This would be because the average of the ping loss during that 3 hour period was < 10%.