Published at: 2026-09-17

How to configure data auditing

Periodically detect data inconsistencies and business anomalies.

Overview

Data auditing batch-checks data inconsistencies across the system. When implementing a CRM system, enterprises typically configure multiple preset and custom objects. Business constraints exist between objects and their fields — for example, order amount equals the sum of order line-item amounts. If business rules, functions, workflows, or system error handling deviate from the design, data may not conform to business requirements.

By configuring data auditing rules, the system periodically scans data, executes validation logic, and sends anomaly results to designated personnel. Administrators can use this information to correct anomalous data or improve the business rules that caused the anomalies.

Before you begin

[!IMPORTANT] - Permission requirements: You need Admin Console permissions for data maintenance tools or the data auditing tool. - Rule preparation: Define the business validation criteria first, such as uniqueness, field-level equations, or cross-object aggregation relationships. - Data preparation: Confirm that auditing target objects, fields, relationships, and filter conditions are already configured. - Notification recipients: Identify the personnel who should receive auditing anomaly alerts in advance.

Applicable scenarios

Scenario	Description	Example
Data deduplication check	Checks whether one or more field combinations in a single object are unique.	An order number can only correspond to one order.
Single-object data auditing	Checks consistency among multiple fields within the same object.	For order products: unit price × quantity = subtotal.
Cross-object data auditing	Checks data consistency between two objects, with aggregation support.	Order amount = SUM(order product unit price × quantity).

How it works

The system generates data scanning and comparison SQL based on the configured auditing rules. After rule execution, the system outputs data records that match the anomaly conditions. Administrators can view anomaly details through notification messages or result objects.

[!NOTE] Data auditing detects anomalies but does not directly correct data. Anomaly correction typically requires administrators to use business data pages, import/export, batch processing tools, or business configuration.

Step 1: Open the data auditing tool

Log in to the Admin Console.
Go to Data Maintenance Tools > Data Auditing Tool.
On the data auditing rule list page, click Create.

Step 2: Fill in basic information

In the basic information section, fill in the rule name, execution cycle, execution time, and notification recipients.

Field	Description	Required
Rule name	Name of the data auditing rule. Use a name that reflects the object and validation target.	Yes
Execution cycle	The time range during which the rule is active and executes.	Yes
Execution time	The frequency of rule execution. Supports minutes, hours, days, or Cron expressions.	Yes
Execution time value	The specific execution time value, such as every 15 minutes, every 1 hour, or daily at 23:00.	Yes
Alert recipients	Personnel who receive auditing result notifications.	No
Rule description	The purpose, business criteria, or handling instructions for the rule.	No

[!TIP] Include the object, field, and validation target in the rule name. For example, Order Amount vs. Line-Item Amount Consistency Audit.

Step 3: Configure auditing data

In the Auditing Data section, select the data object to audit.
For single-object auditing, select the target object and field scope.
For cross-object auditing, select two objects and their related fields.
If the two objects have a Lookup relationship, the system auto-populates the related fields.
Configure object filter conditions as needed to narrow the auditing scope.

Common filter condition example:

Created Time = Schedule execution day

[!NOTE] Cross-object auditing is suitable for master-detail, lookup, or other objects with stable relationship fields. If the relationship is unclear, map the object model first.

Step 4: Select the auditing method

Choose the auditing method based on the business question.

Auditing method	Use case	Configuration focus
Data deduplication check	Checks whether data in an object violates uniqueness requirements.	Select uniqueness fields or field combinations.
Same-object data auditing	Checks calculation relationships among fields within the same object.	Configure field calculation expressions and comparison conditions.
Cross-object data auditing	Checks aggregation relationships between a primary object and a related object.	Configure related fields, aggregation method, and comparison conditions.

Step 5: Configure grouping rules

When two objects have a one-to-many relationship, you typically need to configure grouping on the “many” side.

Select the grouping field.
Set post-grouping filter conditions.
Configure the aggregation calculation criteria.
Confirm the comparison relationship between the grouped result and the primary object field.

Example:

Sales order product line items non-empty data group: COUNT(*) > 0

Step 6: Save and enable the rule

Check the rule name, auditing objects, related fields, and filter conditions.
Check the execution cycle and execution time.
Check the alert recipients.
Click Save.
Enable the rule according to the page prompts.

View auditing results

After rule execution, the system sends auditing results to the configured alert recipients via Chats card messages. Administrators can also view received auditing results through the system object Auditing Result Notification.

Error details in auditing results are stored as Excel attachments. Administrators can download the details, locate anomalous data, and handle it according to the business cause.

Handle anomalous data

After discovering anomalies, follow this recommended process:

Download the auditing result details.
Categorize by object, field, and anomaly type.
Determine whether the anomaly originates from historical dirty data, data entry criteria, functions, workflows, import tasks, or business rule defects.
Correct the anomalous data.
If the anomaly stems from business configuration, adjust the corresponding functions, workflows, validation rules, or import rules.
Re-execute or wait for the next auditing cycle to confirm the anomaly is resolved.

Validation method

Configure an auditing rule using a small set of test data.
Manually create a record that matches the anomaly condition.
Wait for the rule to run at the scheduled time.
Check whether the alert recipients received a Chats card message.
Open Auditing Result Notification and confirm the results are visible.
Download the Excel details and confirm the anomalous data matches expectations.

Considerations

Data auditing rules perform batch scans. Start with a narrow data scope and expand gradually.
Avoid excessively high execution frequency. High-frequency rules may impact system resources.
Cross-object auditing depends on stable relationship fields. Incorrect relationships produce inaccurate results.
Auditing results only indicate that data matched anomaly conditions — they do not necessarily mean the data is wrong. Evaluate in the context of the business scenario.
Before putting a rule into production, validate it on a small dataset first.