Optimize for Infrastructure Version 9.6
 —  Guide for Enterprise Transaction Systems  —

Conventions and Definitions for KPIs and Built-In Rules

This document covers the following topics:


Naming Conventions

This section describes the naming conventions for the KPIs that are used for application monitoring and for infrastructure monitoring of Software AG's enterprise products. The administration names and KPI name are listed in the sections KPI Definitions for Application Monitoring and KPI Definitions for Infrastructure Monitoring.

Administration Names

Administration names are used throughout the Optimize administration. They are also displayed in the KPI Instance Detail panel when data for the KPI instance is shown in tabular format (after clicking the View Data button). In general, the administration name of a KPI or dimension is a compound word in which the single words are joined without spaces and are capitalized within the compound ("CamelCase"). For application monitoring, the first part of the administration name reflects the application monitoring scenario. For infrastructure monitoring, the first part reflects the product component.

The following examples show administration names. The scenario or product component name within the administration name is indicated in bold.

Monitor IDs and KPI Names

The monitor ID is displayed in the Optimize monitoring. It denotes an instance of a KPI. For a unique identification, the monitor ID consists of the component's hierarchical dimensions and the KPI name, separated by a dot (.). The single dimensions are also separated by dots. The KPI name is the descriptive name of the KPI containing blanks if appropriate.

The following examples show monitor IDs. The KPI name within the monitor ID is indicated in bold.

Top of page

Aggregation Types

Multiple values obtained by Optimize in a collection interval are aggregated according to the defined Optimize aggregation type. Some KPIs, however, require a special treatment or cannot be used for calculations at all. The following table describes all possible aggregation types.

Type Description
average Optimize calculates the average value of the data collected over each collection interval.
composite The KPI is a composite of two other KPIs by performing a mathematical operation (add, subtract, multiply or divide).

Composite KPIs are not listed for KPI selection in the Monitored Components page of Optimize. Therefore, they cannot explicitly be selected. Composite KPIs are monitored automatically if the corresponding base KPIs have been selected.

delta The value reflects the increase of a counter in the polling interval. For multiple polling intervals in a collection interval, the data points are summarized.
generic The KPI is not used for monitoring. This is generally the type of non-numeric values such as names.

A generic KPI does not have an associated KPI name. In Optimize, a generic KPI can be visualized in the KPI Instance Detail panel by clicking the View Data button.

last value Optimize uses the value of the last data collected in the collection interval.
state Optimize uses the value of the last data collected in each collection interval where the possible values are 1 (online) or 0 (offline).
sum Optimize summarizes the data points collected over each collection interval.

The Type column of the KPI tables in KPI Definitions for Application Monitoring and in KPI Definitions for Infrastructure Monitoring shows the aggregation type that is used for a given KPI.

Top of page

Generally Used Dimensions for Application Monitoring

The following dimensions are used for all application monitoring scenarios:

Level Administration Name Scenario Displayed Value Example
1 ApplicationMonitoring   AppMon-identifier Application Monitoring
2 AppMonScenario   scenario-identifier RPC
3 ApplicationName CICS ECI host-name:port-number daef:1234
IMS Connect host-name:port-number/datastore daef:1234/IMS0F
RPC application-name DAEF_V82_Collector

The AppMon-identifier is always "Application Monitoring".

Possible values for scenario-identifier are "CICS ECI", "IMS Connect" and "RPC".

For error monitoring, the constant text "Error" is appended to the third dimension (ApplicationName).

Top of page

Generally Used Dimensions for Infrastructure Monitoring

The following dimensions are used for all Software AG enterprise products and components:

Level Administration Name Displayed Value Remark Example
1 Host host-name For non-sysplex systems sunnat5z5
sysplex-name_host-name For sysplex systems DAEPLEX_DAEF
sysplex-name For sysplex totals DAEPLEX
2 Product product-name   Natural Buffer Pool
3 ≥ component-name component-instance   QA42_QA42GBP

Note:
For sysplex systems, the PLEXNAME parameter in the Adabas/Natural Data Collector profile allows using the host-name without preceding sysplex-name.

The component-name is the name of the product component without spaces (similar to the administration names of the KPIs).

The component-instance is the name of the instance. If required, additional parts are added to make it unique or easier to read.

Examples:

Component Name Component Instance Description
AdabasServer 00123 PRODDB The database name is added to the DBID to make it easier to read.
NaturalBufferPool QA42_QA42GBP The Natural subsystem ID is put in front of the buffer pool name to make it unique.
EntireXBroker 1971 The Broker port name.

Detailed lists of the used dimensions can be found in KPI Definitions for Infrastructure Monitoring.

Top of page

Automatically Monitored KPIs

Object state KPIs (such as "AdabasState" or "AdabasCollectorState") are automatically monitored once the corresponding component type instance has been selected. When you configure the monitored components, the object state KPIs are not provided for selection. It is not possible to manually add an object state KPI to use for a component type, and it is not possible to remove an object state KPI for a component type.

Top of page

Built-In Rules

The pre-defined rules listed in KPI Definitions for Application Monitoring and in KPI Definitions for Infrastructure Monitoring are added automatically to the Optimize rule list the first time the Application Monitoring Data Collector calls the Analytic Engine or the first time the Infrastructure Data Collector is started in your environment. By default, essential rules which verify the state of a component are enabled, whereas the other rules are disabled. These rules are provided as examples. They may need to be modified to provide the desired results for your environment before you enable the rule. If a specific rule is not relevant for your environment, you may disable or even delete it.

Several rules refer to the state of the monitored component. When the component is inactive, Optimize indicates a rule violation.

Most rules fire as soon as the data arrives in Optimize, that is, at the polling interval. For technical reasons, rules which refer to composite KPIs fire at the accumulation time, that is, at the KPI interval.

In general, Optimize verifies the rule condition only if a value is provided. If the connection to Optimize is down (for example, if the RPC server is not running for infrastructure monitoring), the Adabas and Natural components cannot provide data. From the Optimize view, the state of the Adabas and Natural components is unknown and the corresponding rules do not indicate a rule violation. Indeed, a failing connection from Optimize to the Adabas Data Collector or Natural Data Collector requires a different action as, for example, an inactive Natural buffer pool. To address this situation, special state KPIs are provided with the Adabas and Natural Data Collectors ("AdabasCollectorState" and "NaturalCollectorState"). These state KPIs indicate "offline" when no data is provided from the collector at the polling time. Like other state KPIs, the Collector state KPIs are automatically monitored. It is strongly recommended to keep an eye on the corresponding rules ("Adabas collector not reached" or "Natural collector not reached"). If any of these rules is violated, the connection from Optimize to the Adabas and Natural Data Collectors must be checked and reestablished so that the Adabas and Natural components can be monitored properly.

Top of page