Data quality SLA rules (IBM Knowledge Catalog)

Data quality SLA rules monitor the data quality of critical data elements, for example, data that serves as input to regulatory reporting, and can initiate remediation tasks for quality issues.

Data quality SLA rules apply to data assets in projects and must be activated for each project separately in the metadata enrichment settings. You can directly access the settings or navigate to the settings from an asset's Data quality page. For more information, see Metadata enrichment settings: Data quality SLA rules. Data quality SLA rules are evaluated when data quality checks in context of a metadata enrichment or data quality rules are run.

Depending on the result of the evaluation, SLA rule compliance or violations are reported on the Data quality page of a data asset that is subject to a data quality SLA rule. See Data quality SLA rule compliance and remediation.

A data quality SLA rule consists of a rule condition and an optional action. The rule condition defines the data elements and the quality criteria to monitor. A basic SLA rule applies to data assets as a whole that are selected by name or by assigned business terms. You can create a finer-grained rules by including subconditions to make the rule apply to individual columns with in selected data asset. The quality criteria defines thresholds for quality scores. The scores of all quality checks that were run on a data asset are aggregated for the data asset, its columns, and the data quality dimensions. The thresholds are applied to the aggregated scores. The selected action determines whether only the results of the SLA rule evaluation are reported or whether a remediation workflow is triggered if the data doesn't meet the quality criteria of the rule condition.

For example, you can create a data quality SLA rule that triggers the default remediation workflow if the completeness and validity dimension scores of any column named EMAIL, EMAIL_ADDRESS, or MAILTO in data assets with the assigned business term Customer contact are not 100%. The definition of that rule consists of a nested condition and an action:

In projects where data quality SLA rules are active, this rule runs whenever a data quality check is run on a data asset that has the business term Customer contact assigned and updates the asset and dimension scores.