Manage Checks

Alation Cloud Service Applies to Alation Cloud Service instances of Alation

Alation Data Quality provides a no-code interface for configuring data quality checks across multiple categories. All checks are executed as SQL queries on the data using the Connector’s Query Service and return a status of Pass, Fail, or Error.

A data quality check is a rule applied to a table or column that evaluates whether the data meets an expected condition and within the defined threshold value.

A threshold is the expected value or range of values that a data quality metric must satisfy during a check. If the observed data meets the threshold, the check passes; if it violates the threshold, the check fails.

Managing Checks in SDK Monitors

The process for defining data quality checks is identical for both Manual and SDK-Enabled monitors. You continue to author, configure, and govern all checks centrally within the Alation UI.

Regardless of the monitor type, the actual query execution follows a pushdown model, meaning the SQL always runs directly on your database (for example, Snowflake, Databricks). The difference lies in how these checks are triggered:

• For Manual Monitors: Alation’s internal scheduler orchestrates the process, triggering the pushdown queries automatically based on the schedule you define in the UI.

• For SDK-Enabled Monitors: Your external pipeline orchestrates the process. The SDK runs within your orchestration environment (such as Apache Airflow or GitHub Actions), where it fetches the check definitions from Alation and triggers the execution of the queries on your target database.

Comparison Operators

Operator	Meaning
=	Equal to
<	Less than
>	Greater than
<=	Less than or equal to
>=	Greater than or equal to
!=, <>	Not equal to
between	Value is within a specified range
not between	Value is outside a specified range

Each check is designed to detect specific types of data quality issues such as missing values, invalid formats, duplicate records, or outdated timestamps.

Result	Meaning
Pass	The data meets the check’s condition
Fail	The data violates the check’s condition
Error	The check failed to execute due to a syntax or runtime issue

The data quality checks are grouped into two main categories:

• Table-level check: Includes numerical, custom SQL query, schema reconciliation, and content reconciliation checks.

• Column-level check: Includes numerical, uniqueness, completeness, validity, custom (common table expressions and SQL query), and metric reconciliation checks.

Data Quality Check Types by Column Data Type

Check Category & Metric	Numerical	Text	Time/Date	Table Level
ACCURACY CHECKS
Average (avg)	Yes	No	No	No
Average Length (avg_length)	No	Yes	No	No
Maximum (max)	Yes	No	No	No
Minimum (min)	Yes	No	No	No
Maximum Length (max_length)	No	Yes	No	No
Minimum Length (min_length)	No	Yes	No	No
Percentile (percentile)	Yes	No	No	No
Standard Deviation (stddev)	Yes	No	No	No
Standard Deviation Population (stddev_pop)	Yes	No	No	No
Standard Deviation Sample (stddev_samp)	Yes	No	No	No
Sum (sum)	Yes	No	No	No
Variance (variance)	Yes	No	Yes	No
Variance Population (var_pop)	Yes	No	Yes	No
Variance Sample (var_samp)	Yes	No	Yes	No
Row Count (row_count)	No	No	No	Yes
UNIQUENESS CHECKS
Duplicate Count (duplicate_count)	Yes	Yes	Yes	No
Duplicate Percentage (duplicate_percent)	Yes	Yes	Yes	No
COMPLETENESS CHECKS
Missing Count (missing_count)	Yes	Yes	Yes	No
Missing Percentage (missing_percent)	Yes	Yes	Yes	No
VALIDITY CHECKS
Invalid Count (invalid_count)	Yes	Yes	Yes	No
Invalid Percentage (invalid_percent)	Yes	Yes	Yes	No
TIMELINESS CHECKS
Freshness (freshness)	No	No	Yes	No
CUSTOM CHECKS
Common Table Expression (cte)	Yes	Yes	Yes	No
SQL Query (sql)	Yes	Yes	Yes	Yes
RECONCILIATION CHECKS
Metric Reconciliation (metric_reconciliation)	Yes	Yes	Yes	Yes
Schema Reconciliation (schema_reconciliation)	No	No	No	Yes
Content Reconciliation (content_reconciliation)	No	No	No	Yes