Health Indicators and KPIs

Health indicators (HIs) provide fine-grained measurements on the CIs that represent your monitored applications and business services. Some HIs provides business metrics such as backlog and volume, while others monitor various aspects of performance and availability such as CPU load or disk space.

There are two types of data sources that can contribute to an HI's status and value: events, and metrics. Some data collectors such as SiteScope send events to Service Health (for example, CPU load exceeded threshold), while other such as Real User Monitor send samples containing metrics (for example, response time = 6 milliseconds).

When an event is sent to Service Health, it is sent with an ETI (event type indicator). The ETI includes a name and a state, for example CPU_Load:exceeded. Using HI definitions in the indicator repository, Service Health translates the ETI state into one of the standard Service Health statuses (Critical, Major, Minor, and so on).

Metric-based HIs apply calculation rules to the samples generated by the data collectors, to create a calculated HI value. For example, Business Process Insight can collect several response time samples over a 15 minute period. A calculation rule will calculate the average of all those samples, and set the HI's status and value accordingly.

The indicator repository enables you to define that when the status of a specific metric-based HI changes, an event is generated. This event then appears in the Event Browser, showing that the HI status has changed.

KPIs are high-level indicators of CI performance and availability, which apply calculation rules to the data provided by HIs to determine CI status. KPIs can be calculated using statuses of HIs, KPIs, or a combination of these. For example, you can specify a rule that sets the severity of the KPI to the worst severity status of any assigned HI, or to the average severity status of all child KPIs.

The value that results from the calculation is used to set a severity level for the KPI based on the KPI definitions; KPI severity can be normal, warning, minor, major, or critical. The resulting measurement for the KPI is translated into a color-coded status indicator displayed in Service Health, where the color represents a more desirable or less desirable condition for the KPI.

You can define a KPI to only use specific HIs that are of interest to you. For example, the BPI Backlog KPI has two HIs: Backlog Value and Backlog Count. If you are interested only in the financial aspects, you can set the KPI to only include the Backlog Value HI in its calculation.

Domains are groups of KPIs that monitor similar functions; for example, the Network domain contains the Network Availability and Network Performance KPIs, while the Business Health and Business Impact KPIs are assigned to the Business domain.

The following diagram illustrates the relationship between KPI domains, CITs, and indicators. (Note that this diagram shows typical indicators for each domain. For the full list of indicators on each CIT, access the indicator repository.)

In this example, you can see that the Business domain includes KPIs and HIs that monitor Business Service and Business Process CITs. The data collector monitoring these CIs is Business Process Insight. The HIs monitored for these CITs include Number of Failures, BP Step Backlog, and Value of Payments. These HIs are used to calculate the following KPIs: Delays, Value, Volume, and Backlog.