Skip to main content
When you create a metric, you select a data store, table, and metric type. Some metric types also require a column. These selections define the metric’s scope by mapping the data quality check to the table or field it evaluates. Data Observability runs the metric against the selected database or warehouse at its configured frequency. The monitored data can be produced by an Integrate.io pipeline, but Data Observability can also monitor supported data stores independently.
Metrics list page showing configured metrics

Metric detail view with data store, table, and column information

Metrics Settings

You can update Metrics name, Data Store, Table name, Column name, and Metric Type. Adding or deleting an existing condition is also possible. Each metric type has a level, listed in Available Metrics Types:
  • Table-level metrics run against the selected table. Some, such as Count, evaluate the whole table. Others, such as Nulls, Cardinality, Duplicate, Min, Max, and Freshness, evaluate a selected column in that table.
  • Pipeline-level metrics, such as Sync Lag and Replication Delay, use pipeline-specific configuration and have separate availability requirements.
The fields shown in the metric form can change according to the selected metric type. For example, a Nulls metric requires a column, while a Count metric evaluates the selected table. You can also add conditions to limit which rows are included in the metric calculation. Conditions can be combined using AND or OR logic and organized into groups. See Available Metrics Types for each metric’s level, supported data types, calculation, and availability.
Metrics settings form with name, data store, and table fields

Metrics settings with column, metric type, and conditions
Changes to a metric’s details apply to every alert configuration that uses the metric as a trigger. Review the affected alerts before saving your changes.

Map a metric to the data it monitors

For example, you can map a Nulls metric to a customer email column or map a Count metric to an orders table. Each selection determines where Integrate.io runs the data quality check:
  1. Data Store identifies the database or warehouse connection.
  2. Table name identifies the table containing the data.
  3. Column name, when required, identifies the field the metric evaluates.
  4. Metric Type defines the calculation to perform.
  5. Conditions optionally restrict the rows included in that calculation.
To change what an existing metric evaluates, edit its Metric Details and select a different available data store, table, column, metric type, or condition. Before saving, review every alert that uses the metric because the changes apply to all of those alert configurations. If an expected table or column is unavailable, confirm that:
  • The table or column exists in the selected database or warehouse.
  • The Data Observability connection has permission to access it.
  • The selected metric type supports the column’s data type.
  • The correct data store and table have been selected.

Archiving a Metric

Click the vertical-ellipsis (⋮) then click Archive metric
Context menu with Archive metric option
Option to Unarchive a Metric is also available.
Context menu with Unarchive metric option
Last modified on October 27, 2026