KPI composer

KPI Composer helps administrators define, compute, and visualize new KPIs from a variety of data sources in real-time. This enables slicing of real-time signaling and traffic telemetry across many dimensions and computes unique KPIs relevant to network operations, engineering, and analytics.

You can use the KPI composer page in the System menu to configure a custom KPI. After the KPI is configured and saved, it computes in real-time and can be visualized in the Composed KPI tab in the Explore page.

These topics describe the various options available to configure a custom KPI.

Name and description

Enter a name and optional description for the new KPI definition. The name field supports alphanumeric characters, underscores and spaces.

  • Valid KPI names - deriv_mcs, kpi composer KPI, kpi 51
  • Invalid KPI names - some-kpi, $foo#kpi12, invalid ! character

The description field supports arbitrary characters.

Preset filters

Select any preset filters to apply to the KPIs. Presets must be defined prior to use. See Presets for details on how presets are configured.

Event sources, categories and metrics

KPI Composer enables KPI creation using events from the following data sources as input.

LTE CTR

All events present in the LTE CTR data feed. These event categories and their metrics follow the same naming conventions as the Ericsson LTE CTR specification.

NR CTR

All events present in the NR CTR data feed. These event categories and their metrics follow the same naming conventions as the Ericsson NR CTR specification.

Collated Records

RSPE records created by joining events from different data streams belonging to the same subscriber session.

Event filter

The event filter is used for filtering out events based on certain criteria. Filters can be applied to any parameters/metrics of the event, and must evaluate to true or false. The event is used for computation only if the filter evaluates to true.

All metrics used in the event filter must first be selected from the Available Metrics section. The variable names used in the filter (like V0, V1 etc) must match the variable names in the Selected Metrics section.

Standard operators

  • Equals == - applicable to all types of metrics
  • Not Equals != - applicable to all types of metrics
  • Greater than > - only applicable to numerical metrics
  • Greater than or equals >= - only applicable to numerical metrics
  • Less than < - only applicable to numerical metrics
  • Less than or equals <= - only applicable to numerical metrics

Special operators

  • IN operator - Check if a metric belongs to a specified set of values. For example, V0 IN (EMERGENCY, HIGH_PRIORITY_ACCESS, MO_SIGNALLING) OR V1 IN (1, 5, 10, 12).

Logical operators

  • AND - Signifies that both conditions must be satisfied in order to evaluate to true. For example, A == 1 AND B > 2.
  • OR - Signifies that at least one of the conditions must be satisfied in order to evaluate to true. For example, A == 1 OR B > 2.
  • ! - Signifies a logical not, or inversion operation. Inverts the result of the expression inside parentheses. For example, !(A < 1 AND B == 5).

An arbitrary number of conditions can be applied to any KPI using the operators defined above.

Expression

The expression defines how the KPI is calculated from the selected metrics.

Similar to the event filter, all metrics used in the expression must first be selected from the Available Metrics section.

Expressions are evaluated using standard operator precedence (BODMAS rule). The expression field accepts arbitrary arithmetic expressions created using basic arithmetic operators and a few special functions.

Arithmetic Operators

  • Addition
  • Subtraction
  • Multiplication
  • Float Division - /
  • Integer Division - //
  • Parentheses

Functions

  • Exponent/Power function - pow(EVENT_PARAM_BASE, EVENT_PARAM_EXPONENT)
  • Natural (base e) log - log(EVENT_PARAM_FOO)
  • Base 10 log - log10(EVENT_PARAM_FOO)
  • Log function custom base - log(EVENT_PARAM_VALUE, EVENT_PARAM_BASE)
  • Convert a number to a string - toString(EVENT_PARAM_TIMESTAMP_START)

Aggregation configuration

This section defines how the KPI is aggregated over time and visualized in the Composed KPIs tab in the Explore page.

The following aggregation types are supported.

  • none - Do not aggregate values over time, preserve raw KPIs.
  • max - Select the maximum value within the selected time interval.
  • min - Select the minimum value within the selected time interval.
  • mean - Compute a mean of all values within the selected time interval.
  • sum - Add all values over the selected time interval.
  • histogram_left - Group values into bins based on provided histogram boundaries. Values coinciding with bin boundaries will be put into the left bin.
  • histogram_right - Group values into bins based on provided histogram boundaries. Values coinciding with bin boundaries will be put into the right bin.

Output Units

Set the output units for the composed KPI. This field is for user reference only and its choice does not affect the computed KPI. The options available for selection are

  • none
  • bytes
  • bps (bits per second)
  • dB (decibels)
  • dBm (decibel-milliwatts)
  • msecs (milliseconds)
  • percentage
check-circle-line exclamation-circle-line close-line
Scroll to top icon