Previous Topic: ADD/FAMILYNext Topic: ADD/WORKLOAD


ADD/HISTORY

This command adds a history file descriptor to the Performance Management parameters file.

Format

ADD/HISTORY name

Parameter

name

The history file descriptor name is a 1 to 20-character alphanumeric string. Use this name with the /HISTORY qualifier to generate reports, graphs, and models from historical data.

Description

The ADD/HISTORY command adds a history file descriptor to the Performance Management parameters file. A history file descriptor defines how the Performance Agent software summarizes daily data. You can have multiple history file descriptors. History file descriptors defined in the parameters file control the archival process for daily Performance data. For a discussion on how to archive data, see ADVISE ARCHIVE.

Performance Agent summarizes process data in one of two ways: unclassified, as defined with /MODEL_DATA, or by the workload families specified with the /CLASSIFY_BY qualifier. When you specify /MODEL_DATA, the archived data retains enough process-level detail to defer the classification until the archived data is used, rather than classifying it when it is archived.

If you omit both /MODEL_DATA and /CLASSIFY_BY history file descriptor elements, Performance Agent archives and reports on data based on interactive, batch, and network workloads.

Qualifiers

/CLASSIFY_BY=USERGROUP=(workload-family[,. . .])

Specifies that Performance Agent store process data in the history file, summarized by the workload families indicated. The /CLASSIFY_BY qualifier cannot be used with the /MODEL_DATA qualifier.

After Performance Agent archives data with a given classification scheme, all subsequent reporting must use one of the classification workload families specified.

/DELETE_AFTER=number_of_days

Specifies the retention period for a history file, in days. The default is 180 days.

/GRANULARITY={DAILY | WEEKLY | MONTHLY | QUARTERLY | YEARLY}

Specifies the time span represented in an individual history file. The default value is monthly.

Granularity

Definition

DAILY

Hours 00:00 to 23:59

WEEKLY

Monday through Sunday

MONTHLY

The first day through the last day of the month

QUARTERLY

January through March, April through June, July through September, October through December

YEARLY

January 1 through December 31

/HOLIDAY_SCHEDULE=(day=(hour-range)[,. . .]|
NOday[,. . .])

Selects a part of or all of the daily data for holidays to be stored in the history file.

The template holiday file PSDC$EXAMPLES:PSDC$HOLIDAYS.TXT enables your holiday schedule. You can edit and move this file to the PSDC$DATABASE area. When using the /HOLIDAY_SCHEDULE qualifier, use the following parameters:

day

Set as SUNDAY, MONDAY, TUESDAY, WEDNESDAY, THURSDAY, FRIDAY, SATURDAY, EVERYDAY, WEEKDAYS, or WEEKENDS.

hour-range

Is specified as m-n, where m and n are numbers from 0 to 24, and m is less than n. More than one hour range can be specified for a given day. Hour-range cannot be used with the NOday option.

/INTERVAL=minutes

Specifies the time period over which Performance Agent averages daily data records into a single history data record in the history file. The default is 60 minutes. The interval must be 2, 4, 6, 8, 10, 12, 16, 18, 20, 24, 30, 32, 36, 40, 48, 60, 72, 80, 90, 96, 120, 144, 160, 180, 240, 288, 360, 480, 720, or 1440 for Performance Management data.

When larger interval values are specified, less disk space is used by the history file. However, this also causes performance data to be summarized over longer periods of time, reducing peaks and valleys of resource use in follow-on processing.

/MODEL_DATA

Indicates that modeling data is to be saved in the history file. The /MODEL_DATA qualifier allows you to generate a baseline model from this history data. This qualifier cannot be used with the /CLASSIFY_BY qualifier. By default, the /MODEL_DATA qualifier is disabled.

History files generated when the /MODEL_DATA qualifier is enabled provide you total and unrestricted ability to generate subsequent reports or graphs with any existing or future classification scheme defined in the parameters file.

/PERIODICITY ={DAILY | WEEKLY | MONTHLY | QUARTERLY}
/NOPERIODICITY

Specifies the calendar period into which the daily files are summarized. You can summarize time spans of up to a year into an average typical day, week, month, or quarter.

You can summarize a week into a typical day; a month into a typical day or week; a quarter into a typical day, week, or month; and a year into a typical day, week, month or quarter.

Consider using graph averaging in place of periodicity. Graph averaging requires more CPU time while periodicity requires more disk space. For more information on graph averaging, see the Performance Manager Administrator Guide.

Valid periodicity values depend upon the selected granularity. The value of the periodicity must be less than the value of granularity. The default is NOPERIODICITY.

Periodicity

Definition

DAILY

Average day

WEEKLY

7 average days

MONTHLY

31 average days

QUARTERLY

93 average days

The following table shows the effect of periodicity on a file defined with yearly granularity:

Periodicity

Definition

DAILY

Averages each daily record into the same typical day. Performance Management archived data is summarized into one typical day record.

WEEKLY

Averages daily data into a typical week for a year. Performance Management archived data represents an average week in the year.

MONTHLY

Averages daily data into a typical month for a year. Performance Management archived data represents an average month in the year.

QUARTERLY

Averages daily data into a typical quarter for a year. Performance Management archived data represents an average quarter in the year.

/SCHEDULE=({day=(hour-range)[,. . .]|Noday}[,. . .])

Selects the daily data to be stored in the history file. You can select all or just some portion of the daily data. When using the /SCHEDULE qualifier:

day

Set as SUNDAY, MONDAY, TUESDAY, WEDNESDAY, THURSDAY, FRIDAY, SATURDAY, EVERYDAY, WEEKDAYS, or WEEKENDS.

hour-range

Is specified as m-n, where m and n are numbers from 0 to 24, and m is less than n. More than one hour range can be specified for a given day. Hour-range cannot be used with the NOday option.

Example

PSDC-EDIT> ADD/HISTORY/SCHEDULE=(WEEKDAYS=13-15,NOWEEKENDS)/MODEL_DATA PEAKLOAD

This example shows you how to create a new history file descriptor named PEAKLOAD.