Aggregations

An aggregation is a value that results from a mathematical operation that calculates metrics from field values. Business Performance Center supports the following aggregation types: Average, Count, Unique count, Max, Min, Sum, Variance, Standard deviation and Latest.

The following table describes the operation of the various aggregations supported by Business Performance Center.

Table 1. Aggregations purpose
Aggregation type What it calculates
Average Arithmetic mean of all numeric field values.
Count Total number of events or records.
Unique count Number of distinct values in a field.
Max Highest numeric value among the field values.
Min Lowest numeric value among the field values.
Sum Total sum of numeric field values.
Variance Degree of variability in numeric field values, showing how much the values differ from the average.
Standard deviation Square root of the variance that indicates the level of dispersion in numeric field values and how consistently or widely the values vary around the average.
Latest Value of the field from the most recent event (according to timestamp/order)
Note: Because Business Performance Center uses OpenSearch as its database, approximation problems can be found with aggregations of numbers greater than 2^53 (that is, about 9 * 10^15). See the Limitations External link opens a new window or tab section of the Aggregations page of the OpenSearch documentation. Additionally, approximation problems can be found with Unique count aggregations of numbers greater than 3000 by default. See the Cardinality aggregations External link opens a new window or tab page of the OpenSearch documentation.

Variance and Standard deviation aggregations

The Variance and Standard deviation aggregations provide statistical measures of data dispersion. They help you understand how much numeric values vary from their average, providing additional analytical context beyond basic aggregations such as Sum or Average.

These aggregations are particularly useful for analytics-driven dashboards, risk assessment, anomaly detection, and data quality validation.

  • Variance measures how far each value in a dataset deviates from the mean.
  • Standard deviation is the square root of the variance and expresses dispersion in the same unit as the original data.

A low value indicates that data points are close to the mean. A high value indicates greater variability.

A high value indicates that the data points are spread out (high variability).

Depending on your use case, high variability may indicate:

  • Risk (for example, fluctuating financial metrics)
  • Outliers or anomalies (for example, cybersecurity threats)
  • Data quality issues (for example, manipulated or inconsistent data)

In other scenarios, variability may indicate growth or dynamic behavior (for example, increasing revenue spread).

These aggregations can be applied to numeric fields only, including:

  • Integer
  • Long
  • Float
  • Double

If the selected data item is not numeric, the aggregation option is not available.

Grouping behavior

When a Group by field is configured:

  • The aggregation is calculated independently for each group.
  • Results are displayed per category (for example, per country).
  • The computation is performed on the subset of records belonging to each group.

This allows comparison of variability across segments of your data.

Handling of null and missing values
  • Null or missing numeric values are ignored during calculation.
  • Only valid numeric records are included in the computation.
  • If no valid numeric records exist for the selected field, no value is returned.

Performance considerations

Standard Deviation and Variance are computed by the OpenSearch backend and are optimized for large datasets. However, performance depends on:

  • Total number of records
  • Number of groups (when using Group by)
  • Field cardinality

For best performance, avoid grouping on high-cardinality fields unless necessary and ensure appropriate indexing for numeric fields.

Example use cases

These aggregations extend reporting capabilities from simple descriptive metrics to more advanced analytical insights, supporting compliance and analytics-focused requirements.

  • Financial risk analysis: Assess variability in loan repayments or income.
  • Operational monitoring: Detect abnormal fluctuations in processing times.
  • Cybersecurity analytics: Identify extreme deviations that may indicate threats.
  • Data quality monitoring: Detect unusual variations in feature distributions before applying machine learning or AI models.

Latest aggregation

The Latest aggregation returns the most recent value of a selected numeric field based on timestamp ordering. This aggregation is useful when only the most up-to-date value is relevant, such as the current price of a product or the most recent status of a business activity.

Key features

  • Timestamp-based sorting: The aggregation identifies the most recent record by sorting data using a timestamp field.
  • Customizable date field: In Periodic Metric and Periodic KPI measurements, you can select which date field to use, such as date created, date last modified, or date checked in.
  • Default timestamp behavior: In Metric measurements, the system automatically uses the default timestamp field.
  • Numeric fields only: The Latest aggregation applies only to numeric fields like age, price, or quantity. It does not support text or keyword fields.
  • Monitoring source dependent: The list of available fields for aggregation depends on the selected monitoring source. Users will see different field names depending on the structure of the source data.
  • Multiple aggregations: Users can apply the Latest aggregation alongside other aggregations like Count or Sum in the same chart, provided they are applied to different fields.

Unlike other aggregation types, the Latest aggregation requires a timestamp field to determine ordering. Aggregations such as Sum, Count, or Average are calculated independently of time.

How to use the Latest aggregation

  1. Create a Chart.

    Start by creating a chart using either the Metric or Periodic Metric measurement type.

  2. Select the Aggregation Type.

    In the chart configuration panel, click Add Aggregation and choose Latest from the list.

  3. Select a numeric field such as age, price, or quantity. Only number-type fields are supported for Latest aggregation.

    You can apply as many aggregations as you want in the same field.

  4. If you are using a Periodic Metric chart, select the timestamp field that determines which record is considered the most recent. Available fields depend on the selected monitoring source.

    In Metric charts, the default timestamp field is used automatically and cannot be changed.

  5. Optional: Apply Group By.

    You can group the data by keyword or boolean fields to see the latest value per category. For example, the latest price per product or the latest age per user.

  6. Preview and Save.

    Use the Table chart type to verify the values returned by the Latest aggregation. This is especially helpful during testing or when comparing against raw data.