Metrics

Metric

A metric, also known as a measure, is a standard for measuring change or development. Examples include temperature, population, GDP, revenue, number of users, profit margin, retention rate, and coverage rate.

Metric domain

A metric domain is the object associated with a metric, used to categorize and manage metrics. IoT Data Service provides two metric domains: product and device.

Metric type

• Original definition

  Original definition metrics are derived from the thing model of a product in IoT Platform.

  In the IoT Platform console, navigate to Device Management > Product, select a product to open its Product Details page, then click Function Definition to view the defined thing model.

• Raw metric

  Raw metrics originate from properties that devices report to IoT Platform or from properties parsed from custom topics.

  In the IoT Platform console, navigate to Device Management > Device to open the Device Details page. Click TSL Data to view properties. For example, on the Device Details page, choose TSL Data > Running Status to view properties reported by a device, such as metrics from an electrical parameter measuring instrument: load power (28 kW), operating current (15 A), and voltage (20 V).

• Derived definition

  A derived definition is based on an original definition and applied to the sub-entities of a product.

  Similar to an original definition, a derived definition is a property definition and does not have a specific data value.

  The aggregated data becomes the value for each device's derived metric.

  A product's derived definition and the derived metric for each device under that product share the same metric code.

• Derived metric

  A derived metric is calculated from raw metrics, original definitions, or derived definitions by using aggregation methods such as sum and average.

  For a derived metric, a time modifier defines the aggregation time range (the time constraint), and business modifiers define other constraints.

  Therefore, a derived metric consists of three elements: time modifier (required) + business modifier (optional) + atomic metric (required).

  To learn how to add a metric, see Add Metric.

Metric type applicability

The following table shows which metric types are supported in each metric domain.

A device under a product can include metrics of both the derived definition and derived metric types.

For example, consider a derived definition for product PK1 named Average Temperature in the Last 24 Hours. This product has 100 devices, named DNn (where n=1 to 100). Creating a separate Average Temperature in the Last 24 Hours derived metric for each device would require repeating the configuration 100 times. However, creating a single derived definition for product PK1 automatically creates the Average Temperature in the Last 24 Hours derived metric for all 100 devices.