P. Glossary

action

An action executed automatically by Performance Management when the data being monitored reaches a threshold value. The following actions are supported:

Sending an email
Executing a command
Issuing an SNMP trap
Issuing a JP1 event

Action Handler

A PFM - Manager or PFM - Base service that executes actions.

Agent

A PFM - Agent service that collects performance data.

Agent Collector

A PFM - Agent service that collects performance data and evaluates the data according to the threshold values set in alarms.

Agent Store

A PFM - Agent service that stores performance data in a database. The Agent Store service uses a database for recording performance data. Each PFM - Agent has its own Agent Store service.

alarm

Information that defines an action or event message that is triggered when the data being monitored reaches a threshold value.

alarm table

A table containing the following definition information about one or more alarms:

Monitored object (process, TCP, Web service, and so on)
Monitored information (CPU usage, number of bytes received per second, and so on)
Monitored condition (threshold value)

binding

The process of associating alarms with an agent. Binding enables the user to be notified when the performance data collected by the agent reaches a threshold value defined in an alarm.

cluster system

A single system configured from multiple linked server systems. There are two major types of cluster systems: an HA (High Availability) cluster system and a load-balancing cluster system.

In this manual, a cluster system means an HA cluster system.

[Figure] HA cluster system

[Figure] Load-balancing cluster system

Correlator

A PFM - Manager service that controls event distribution between services. This service evaluates the alarm status, and sends an alarm event or agent event to the Trap Generator service and to PFM - Web Console if the alarm status exceeds a threshold value.

database ID

An ID attached to each record in PFM - Agent, indicating the database in which the record is stored and the record type. The database ID may be either of the following:

PI

Indicates that the database contains records of the PI record type.
PD

Indicates that the database contains records of the PD record type.

data model

A generic term for the records and fields contained in a PFM - Agent. Data models are versioned.

drilldown report

A report related to another report or to the fields in the report. A drilldown report can be used to display detailed information or related information for a report.

executing node

Of the server systems in a cluster system, the node that is currently executing applications (node whose logical host is active).

failover

The process by which the standby node takes over processing if a failure occurs on the node that is executing applications in a cluster system.

field

Individual operation information entries in a record. Each field serves as a monitoring item for Performance Management.

Function ID

A one-byte identifier indicating the function type of a service of Performance Management programs. This is part of the service ID.

HA cluster system

A cluster system designed to implement high availability by continuing operation even if one system fails. If a failure occurs on the server currently executing applications, a separate standby server takes over and continues the processing of applications. Accordingly, because application processing is not interrupted when a failure occurs, availability improves.

In this manual, a cluster system means an HA cluster system.

historical report

A report that tracks the status of an object being monitored from a point in the past to the present.

instance

In this manual, the term instance is used as follows.

To indicate the format of a record:

A record written on one line is known as a single-instance record. A record spanning multiple lines is known as a multi-instance record, each line of which is known as an instance.
To indicate the number of PFM - Agent:

A single agent that monitors all the target objects on a host is known as a single-instance agent. Agents that share the monitoring of target objects on a host are known collectively as a multi-instance agent. Each of these agent services of a multi-instance agent is called an instance.

instance number

An identifier for management number used for internal processing. An instance number is part of the service ID.

lifetime

The length of time that the consistency of the performance data collected in each record is retained.

JP1/SLM

A product that performs monitoring from the viewpoint of performance as experienced by the service users of a business system, and that supports service-level maintenance. Linkage with JP1/SLM can enhance monitoring of the operating status.

load-balancing cluster system

A system that distributes the processing load over multiple nodes to improve throughput. Because processing switches to another node if an executing node stops due to a failure, this system also improves the availability of the system.

logical host

A logical server that provides the JP1 execution environment for operation in a cluster system. If a failure occurs on the executing node, the logical host is switched to the standby node. Each logical host has a unique IP address. At failover, the IP address is inherited by the standby node. Thus, when the physical server is failed over, clients can still access the logical host using the same IP address. To the clients, it appears that one server is operating continuously.

management tool

Any command or GUI-based function used to verify the status of a service or to manipulate performance data. Management tools allow you to:

Display the configuration and status of a service
Save and restore performance data
Export performance data to a text file
Delete performance data

Master Manager

A PFM - Manager service. This is the main service of PFM - Manager.

Master Store

A PFM - Manager service that manages the alarm events issued from each PFM - Agent. This service uses a database to store the event data.

monitoring template

A set of predefined alarms and reports provided by PFM - Agent. The monitoring template facilitates preparation for monitoring of the PFM - Agent operation status without the user having to enter complex definitions.

multi-instance record

A record spanning multiple lines. This type of record has unique ODBC key fields.

[Figure] instance

ODBC key field

These fields display the primary keys that are necessary to use the data retrieved from records stored in the Store database on either PFM - Manager or PFM - Base. Some ODBC key fields are common to all records; others are record-specific.

PD record type

[Figure] Product Detail record type

performance data

Data about the operation status of a resource, collected from the system being monitored.

Performance Management

A generic term for a family of software products used to monitor and analyze problems related to system performance. Performance Management consists of the following five program products:

PFM - Manager
PFM - Web Console
PFM - Base
PFM - Agent
PFM - RM

PFM - Agent

One of the program products in the Performance Management family. PFM - Agent is responsible for system monitoring. Several types of PFM - Agent are available, depending on the applications, database, and OS to be monitored. PFM - Agent provides the following features:

Performance monitoring of target objects
Collection and recording of data from monitored objects

PFM - Base

One of the program products in the Performance Management family. PFM - Base provides the core functionality for operation monitoring in Performance Management. It is a prerequisite product for running PFM - Agent and provides the following features:

Commands and other management tools
Common functions for linking Performance Management with another system

PFM - Manager

One of the program products in the Performance Management family. PFM - Manager performs supervisory functions and provides the following features:

Management of the Performance Management program products
Event management

PFM - Manager name

A field name that identifies the field in the Store database that stores the reference data. Use this name, for example, when you execute any command with the field name in the Store database.

PFM - View name

Alias name for PFM - Manager name. PFM - View name is more intuitive than PFM - Manager name. For example, "INPUT_RECORD_TYPE" (PFM - Manager name) is "Record Type"(PFM - View name). Use this field name, for example, when you specify the field name in the PFM - Web Console windows.

PFM - Web Console

One of the program products in the Performance Management family. PFM - Web Console operates as a Web application server to enable centralized monitoring of the Performance Management system via a browser. It provides the following features:

Display in a graphical user interface
Integrated monitoring and management
Definition of reports and alarms

physical host

An environment unique to each server in a cluster system. When a failover occurs, the environment of the physical host is not inherited by the other server.

PI record type

[Figure] Product Interval record type

Product Detail record type

A type of record for storing performance data indicating the system status at a specific point in time, such as detail information about the currently running process. PD records can be used to acquire system statuses such as the following at a specific point in time:

System operation status
Amount of file system capacity currently in use

product ID

A one-byte ID indicating the Performance Management program product to which the service of the Performance Management program belongs. A product ID is part of the service ID.

Product Interval record type

A type of record for storing performance data at set intervals, such as a process count every minute. PI records can be used to analyze such time-based changes and trends in the system status such as the following:

Number of system calls generated within a set time period
Changes in the amount of file system capacity used

real-time report

A report that shows the current status of an object being monitored.

record

A group of operation information entries categorized by their purpose. A monitoring agent collects operation information from each record. The types of records that can be collected vary depending on the agent program.

report

Information defined for graphical display of the performance data collected by PFM - Agent. The main types of information you can define are as follows:

The records to be displayed in a report
The performance data items to be displayed
The display format of performance data (table, graph, and so on)

service ID

A unique ID assigned to each service of the Performance Management programs. You must specify the service ID when you execute a command to check the Performance Management system configuration or to back up performance data of an agent, for example. The format of the service ID differs depending on the setting of the product name display functionality. For details about the format of the service ID, see the chapter on Performance Management functionalities in the JP1/Performance Management Planning and Configuration Guide.

single-instance record

A record written on a single line. This type of record does not have any unique ODBC key fields.

[Figure] instance

stand-alone mode

A PFM - Agent activated as a stand-alone program. If either of the PFM - Manager services Master Manager or Name Server is disabled due to a failure or another problem, you can still collect performance data by starting PFM - Agent.

standby node

Of the server systems in a cluster system, a node that is waiting to take over applications if the executing node fails.

Store database

A database containing performance data collected by the Agent Collector service.

Non-interactive (command)

Command execution method in which operator input required for command execution are provided by values in option specifications or in definition files.

Executing a command non-interactively saves work when configuring an operation monitoring system and can reduce user workload.

To Page Top