Data quality objects define quality frameworks and checks, and Service Level Agreement (SLA) objects define performance commitments. See Data quality and SLAs for the concepts. This group has six object types: Data quality, Quality dimension, Quality check, SLA, SLA dimension, and SLA property.
Data quality
A data quality definition records how good a data product’s data is. It groups the quality dimensions the product measures itself against, such as completeness or accuracy.
The object type comes from the Open Data Product Specification (ODPS). It belongs to the Quality and SLA group of the Data Product Standard application.
A data quality definition is declarative when written as text, or executable when a tool runs it. It can also name that tool, such as sodacl, montecarlo, or dqops. Publishing it lets consumers compare products on the same quality terms.
For example, a Customer 360 product defines a data quality definition measured by dqops, then attaches completeness and accuracy dimensions.
Data quality attributes
| Attribute | Description |
|---|---|
| Display Title | Name shown for the data quality definition. |
| Definition Type | Form of the definition: declarative or executable. |
| Tool Type | Quality tool: sodacl, montecarlo, dqops, or custom. |
| Tool Version | Version of the tool. |
| Reference URL | Link to the quality definition source. |
| DQ Specification | Rich-text or structured quality specification. |
Quality dimension
A quality dimension is a single quality metric within a data quality definition. It sets one measurable target the product commits to, such as accuracy, completeness, or timeliness.
The object type comes from the Open Data Product Specification (ODPS). It belongs to the Quality and SLA group of the Data Product Standard application, nested under a data quality definition.
Each dimension pairs an objective with a unit — the level promised and how it is measured. The unit is a percentage or a plain number. Add one dimension for every quality aspect the product guarantees.
For example, a Customer 360 product adds a completeness dimension with an objective of 99 and the unit percentage.
Quality dimension attributes
| Attribute | Description |
|---|---|
| Display Title | Name shown for the dimension. |
| Quality Dimension | Dimension measured: accuracy, completeness, conformity, consistency, coverage, timeliness, validity, or uniqueness. |
| Objective | Numeric target for the dimension. |
| Unit | Measurement unit: percentage or number. |
Quality check
A quality check is a concrete, runnable test that confirms whether data meets a quality target. Where a quality dimension states the goal, a check proves it.
The object type comes from the Open Data Contract Standard (ODCS). It belongs to the Quality and SLA group of the Data Product Standard application. A check attaches to a data contract, a schema object, or a schema property, so it runs at the exact level where the data lives.
Each check sets a metric to measure and a pass condition. A severity — info, warning, or error — grades a failing result.
For example, a Customer 360 contract runs a check on the customer_id property. It counts nullValues and raises an error when any appear.
Quality check attributes
| Attribute | Description |
|---|---|
| Check Name | Name of the check. |
| Check Type | How the check is defined: library, sql, custom, or text. |
| Quality Dimension | Dimension assessed, such as accuracy, completeness, or uniqueness. |
| Severity | Result severity: info, warning, or error. |
| Business Impact | Impact category: operational or regulatory. |
| Metric | Metric measured: nullValues, missingValues, invalidValues, duplicateValues, or rowCount. |
| Method | The approach used to run the check (free text). |
| Schedule | When the check runs. |
| Must Be / Must Not Be | Threshold conditions the metric is validated against (equality, range, and comparison variants). |
| SQL Query | Query used for a sql check. |
| Engine | Execution engine: soda, great-expectations, monte-carlo, or dbt. |
| Tags | Free-text tags for the check. |
SLA
An SLA is a Service Level Agreement — the delivery commitments a data product makes to its consumers. It covers availability, freshness, and the support consumers can rely on.
The object type comes from the Open Data Product Specification (ODPS). It belongs to the Quality and SLA group of the Data Product Standard application.
An SLA holds its commitments as SLA dimensions, each with an objective and a unit. It also records support contacts — an email, a phone number, and the hours each is available. Publishing one makes a product’s reliability explicit.
For example, a Customer 360 product publishes an SLA committing to 99.9% uptime. It lists a support email answered during business hours.
SLA attributes
| Attribute | Description |
|---|---|
| Definition Type | Form of the definition: declarative or executable. |
| SLA Type | The category of the SLA (free text). |
| SLA Specification | Rich-text or structured SLA specification. |
| Reference URL | Link to the SLA source. |
| Support Email | Support contact email and its response hours. |
| Support Phone | Support contact phone and its coverage hours. |
| Support Documentation URL | Link to support documentation. |
SLA dimension
An SLA dimension is one metric inside a Service Level Agreement (SLA). It states a single delivery commitment, such as uptime, latency, or response_time, with a target.
The object type comes from the Open Data Product Specification (ODPS). It belongs to the Quality and SLA group of the Data Product Standard application, nested under an SLA.
Each dimension pairs an objective with a unit, such as percent, milliseconds, seconds, or days. Add one dimension for every service level the product guarantees.
For example, a Customer 360 product adds an uptime dimension with an objective of 99.9 and the unit percent.
SLA dimension attributes
| Attribute | Description |
|---|---|
| Check Name | Name of the check. |
| Check Type | How the check is defined: library, sql, custom, or text. |
| Quality Dimension | Dimension assessed, such as accuracy, completeness, or uniqueness. |
| Severity | Result severity: info, warning, or error. |
| Business Impact | Impact category: operational or regulatory. |
| Metric | Metric measured: nullValues, missingValues, invalidValues, duplicateValues, or rowCount. |
| Method | The approach used to run the check (free text). |
| Schedule | When the check runs. |
| Must Be / Must Not Be | Threshold conditions the metric is validated against (equality, range, and comparison variants). |
| SQL Query | Query used for a sql check. |
| Engine | Execution engine: soda, great-expectations, monte-carlo, or dbt. |
| Tags | Free-text tags for the check. |
SLA property
An SLA property records a Service Level Agreement (SLA) commitment on part of a data contract. It binds one measurable promise to a specific element of the contract’s data.
The object type comes from the Open Data Contract Standard (ODCS). It belongs to the Quality and SLA group of the Data Product Standard application.
Each property carries a value, a unit, and the element it applies to. A driver — regulatory, analytics, or operational — records why the commitment exists. A schedule sets when it is evaluated.
For example, a Customer 360 contract adds an SLA property for 99.9% uptime on one schema object. Its driver is operational.
SLA property attributes
| Attribute | Description |
|---|---|
| Property Name | Name of the service level property. |
| Value | Committed value. |
| Unit | Unit the value is expressed in. |
| Element | Schema element the property applies to. |
| Driver | Reason for the commitment: regulatory, analytics, or operational. |
| Schedule | When the property is evaluated. |
For the concepts behind these object types, see Data quality and SLAs. For how they connect to other object types, see Data product object types and attributes.