TimeProfiles Section

The [TimeProfiles] section defines the time intervals that Stat Server references for calculating historical, aggregate values for statistics. This section must be named TimeProfiles within the Stat Server Application object. Stat Server clients, such as CCPulse+, specify which defined time profile to use when they request statistics. The following table lists the one configuration option that is applicable to the [TimeProfiles] section.

Configuration Option for the TimeProfiles Section

Option

Description

<TimeProfileName>,<Type>

Defines the time interval over which a historical aggregate value is calculated. The option name must consist of two entries separated by a comma: <TimeProfileName> represents any string that names the time profile, and <Type> represents the time interval type, which includes one of the following: Sliding Growing Selection (for Stat Server applications that operate in regular mode only) SinceLogin With the exception of SinceLogin, you must specify values for each interval type. Stat Server uses a special time profile, called Default, if a client does not specify a time profile when requesting statistics. The Default time profile uses a Growing interval type and resets statistics to zero (0) every night at midnight. To override the reset time of this inherent time profile, you must add a Default time profile to the [TimeProfiles] section and redefine it as desired. Default Value: No default value. Valid Values: Dependent on interval type. (See the following subsections.) Changes Take Effect: When Stat Server restarts.

Stat Server projects actions and statuses onto time intervals (except for the TotalAdjustedTime and TotalAdjustedNumber statistical categories) on any time profile, as follows:

• Status duration time for a status in progress is included in a statistic, even if the status is not completed.
• Action duration time for an action in progress is not included in a statistic until the action is completed.

Values for Sliding Interval Type

Values for the Sliding interval type use the following format:

  interval:sampling

where

  interval specifies the duration, in seconds, of the reporting interval.
  sampling (optional) specifies the duration, in seconds of the sampling. If the sampling value is not specified, Stat Server uses its default of 10 seconds.

Example. Suppose that you want to set up a time profile (Last10) that always tracks the last 600 seconds of activity, with a sampling taken every 2 seconds.

Example of Sliding Interval Type

To create this time profile, under the [TimeProfiles] section of your Stat Server application, enter Last10,Sliding in the Option Name field and 600:2 in the Option Value field.

Values for Growing Interval Type

Values for the Growing interval type consist of:

• Time to reset statistics to zero.
  The time to reset statistics is in the 24-hour clock format. For example, 00:00 is midnight, 13:00 is 1:00 PM, and so on.
• (Optional) Increment at which to reset statistics.
  The optional increment is also in the 24-hour clock format and is relative to the time to reset statistics to zero.

If no time profile is specified for a statistic requested by any client, Stat Server calculates statistics using the Growing interval type, which re-sets statistics to zero at 00:00 (midnight) unless a time profile named Default in the [TimeProfiles] section specifies a different initialization time. For example, to set Default to reset at 1 AM instead of midnight, enter Default,Growing in the Name field and 01:00 in the Value field.

Important

To specify more than one set of values, separate the sets with commas.

Example. Suppose that you want to set up a time profile (named Shifts) that resets statistics to zero when shifts change at 3:00 AM, 7:00 AM, 11:00 AM, 1:00 PM, 7:00 PM, and 1:00 AM. To do so, enter Shifts,Growing in the Name field and 3:00 +4:00, 13:00 +6:00 in the Value field.

In this example, 3:00 +4:00 is translated as reset to zero at 3:00 AM, reset to zero at 3:00 AM plus 4 hours (7:00 AM), and then reset to zero again at 7:00 AM plus 4 hours (11:00 AM). The setting 13:00 +6:00 is translated as reset to zero at 1:00 PM (or 13:00 on the 24-hour clock), reset to zero at 1:00 PM plus 6 hours (7:00 PM, or 19:00 on the 24-hour clock), and then reset to zero again at 7:00 PM plus 6 hours (1:00 AM).

The Figure below illustrates this example.

Example of Growing Interval Type

Values for Selection Interval Type

The Selection interval type calculates a time interval defined by the end or occurrence of the specified number of actions or statuses. Stat Server in restricted cluster mode logs an error if you specify Selection for the time profile. This interval type does not apply. A Selection interval lasts until the current time, or until the last action or status out of the specified number of actions or statuses has occurred (for instantaneous actions) or ended (for durable actions and statuses). The first time interval starts when Stat Server starts calculating a particular statistic. At a given moment, no more than the specified number of actions or statuses can occur during one Selection interval.

The actions or statuses taken into account are those listed either in the relative mask of the statistical type on which a statistic is based, or in the main mask if no relative mask is specified for the statistical type (see also Statistical Type Sections). The time interval varies depending on the amount of time it takes for the specified actions or statuses to occur.

The value for the Selection interval type must be an integer.

Important

You can specify a relative mask in a statistical type for the purpose of Selection intervals, even if the statistical category on which the type is based does not require a relative mask.

Example. Suppose that you want to set up a time profile (named Last5Calls) that tracks the last five calls. To do so, enter Last5Calls with an interval type of Selection, and 5 in the Value field.

The Figure below illustrates this example. In it, Total Interval 5 is calculated from the end of Action 4 until Current Time. Because no action is in progress at CurrentTime, the interval only includes durations of four actions (5 through 8).

Example of Selection Interval Type

Values for SinceLogin Type

The SinceLogin interval type aggregates statistical data only for agent-object statistics – that is, statistics based on stat types with object type defined as Agent. Stat Server resets such statistics to zero (0) at the moment of agent login. Statistics continue to accumulate as long as the agent is logged into (any) DN. The SinceLogin interval enables statistic requests by agent. This means you can now identify the least-occupied agent, for example, by requesting every agent’s total handling time with SinceLogin interval.

No other parameters are passed with this interval.

[+] Notification Modes

When requesting statistics, clients also specify how often they expect updates on the statistical values. Stat Server, in both regular and restricted cluster mode, sends updates using one of the following notification modes:

• ChangesBased – Stat Server reports the current value whenever a statistical value changes. For time-related statistics, Stat Server reports the current value whenever a statistical value changes and with the specified notification frequency.
• TimeBased – Stat Server reports the current value at the specified notification frequency (for example, every two seconds).
• ResetBased – Stat Server reports the current value right before setting the statistical value to zero (0).
• NoNotification – Stat Server does not report updates or updates are turned off.

Note that you can also request Current statistics with any of the top three notification modes. These statistics do not require any time profile unless requested with ResetBased notification mode, in which case, you must use the Growing time profile. CurrentState statistics cannot be requested with ResetBased notification mode.

[+] Insensitivity

Some Stat Server client applications, such as CCPulse+, specify an insensitivity value to further control the network "chatter" between agent PCs and Stat Server. Insensitivity describes a condition for Stat Server to send updates of statistical values to its clients. An increase in the value of this parameter usually decreases network traffic, but it also reduces reporting accuracy, because values are not updated as frequently. This setting is not visible in Stat Server configuration, but rather, clients pass its value to Stat Server along with each statistic request.

Insensitivity plays no role for reset-based statistics. For time-based or change-based notification mode, Stat Server only reports the recalculated value if the absolute value of the difference between the previous value and the recalculated value or its percentage ratio to recalculated value is at least equal to the number specified by insensitivity.

In addition, Stat Server uses a different algorithm of comparison with insensitivity depending on the data type of the result Stat Server calculates.

• If the result is a floating-point decimal—as is the case for statistics providing custom values, ratios, or averages—Stat Server uses percentages as the measure of comparison of insensitivity between a previous and a recalculated value. Given an Insensitivity setting of 5 for a floating-point statistic, for instance, Stat Server sends the recalculated result to its client only when the absolute value of the difference between the new and the old result is more than 4 percent of the absolute value previously sent. In the same scenario, but with an Insensitivity setting of 1, Stat Server sends the recalculated result when it differs, by any amount, from the value previously sent.

• If the result has a long integer data type—as is the case for statistics measuring time—Stat Server uses the absolute difference in values for comparison. Given an Insensitivity setting of 5 in this case, Stat Server sends the recalculated result to its client when the absolute value of the difference between the new and old result is at least 5 (seconds, usually).

Tip

This algorithm has changed throughout the releases. In 6.1 and prior releases, Stat Server did not use percentages to measure insensitivity.