Published at: 2026-09-17

How to archive historical data

The data archiving feature lets administrators periodically migrate low-frequency historical business data to distributed large objects based on predefined rules. This significantly optimizes primary database storage resources and improves search and runtime performance for the core system under high data volumes.

Overview

Data archiving logic: Data archiving extracts historical records from standard objects, writes them into large object storage, and automatically physically deletes the original data from the standard object after successful migration. This preserves long-term read-only access and traceability for historical records while keeping the primary database lean and improving front-end list page response times.

Typical business scenarios

Scenario 1: Historical transaction and customer data archiving

Background: Over years of operation, enterprises accumulate massive volumes of customer transaction and communication records. These records were frequently accessed when current, but access drops dramatically over time. However, they must be retained for compliance or audit purposes.
Practice: For example, a manufacturing enterprise periodically archives “vehicle sales order” data older than 3 years to large objects, effectively reducing the load on the core CRM database.

Scenario 2: Periodic data backup

Background: CRM system data is critical to business continuity. Enterprises need to periodically back up specific record sets in an isolated, offline manner.
Practice: Archiving is used periodically as a data protection tool, storing snapshots in distributed large object storage for extreme-case data tracing and recovery.

Before you begin

[!IMPORTANT] * Administrative permissions: The administrator must have functional permissions for Admin Console > Object Management > Data Archiving. System administrators and CRM administrators have this permission by default. Other roles must be granted permission through the Admin Console. * Archiving impact: Archiving is an irreversible, high-risk operation. After data is migrated to a large object, the source object data is permanently physically deleted. Confirm backups or test on a small scope before executing.

Configuration guide

Step 1: Manage the archiving rule list

Go to Admin Console > Object Management > Data Archiving. On this page, you can view and manage all created archiving rules:

Edit: Modify information or filter conditions for disabled rules.
Enable/Disable: After enabling a rule, the system automatically executes archiving at the configured frequency. Disabling a rule pauses the archiving schedule.
Delete: Only rules in the “Disabled” state can be deleted. If a rule is enabled, the system blocks deletion with the message “Rule is enabled and cannot be deleted.”

Step 2: Create an archiving rule and configure basic information

On the data archiving list page, click Create in the upper-right corner.

On the configuration page, fill in the basic information:
- Archive name: A custom rule name (for example, Historical Vehicle Sales Order Archive). Must be unique within the tenant.
- API Name: A developer-facing identifier. Must start with an English letter, support letters, numbers, and underscores, and not exceed 38 characters. Custom object format must end with __c.
- Description: Describes the purpose of this archiving rule.

Step 3: Set the archiving object and field mapping

Configure the data source, destination, and one-to-one field mapping relationships.

Source object: Select the standard object to archive (supports both preset and custom objects). Only one source object can be selected per rule, and it cannot be changed after saving.
Target object: Select the large object to store the archived data. Only large objects not bound to other active archiving rules are available. Cannot be changed after saving.
Field settings and mapping:
- Drag or select fields from the standard object to map them one-to-one to large object fields.
- Required mappings: The Primary Attribute and Owner fields of the source object must be explicitly mapped to the corresponding fields in the target large object.

Step 4: Set the archiving data scope

Define which data is moved to the historical store. Two modes are supported:

All data: All historical data under the source object participates in archiving.
Filter by conditions: Only records matching the filter conditions (for example, Sales Order Status equals Delivered AND Created Time before 3 years ago) are archived.

Step 5: Set the archiving frequency

Configure the cycle and specific time for automatic background archiving:

Frequency: Choose from Once (immediate or scheduled), Daily, Weekly, or Monthly.
Execution time: Set the specific time zone and time point. To minimize impact on daily operations, schedule execution during off-peak hours (for example, 02:00).

Key considerations

[!CAUTION] 1. In-progress flow protection: Records with active flows (including approval flows, business flows, and pipeline instances that have not completed) are not archived. The system automatically skips these records to preserve flow integrity. 2. Historical flow cleanup: Records with completed flows can be archived. After data migration, all associated approval trails and flow history in the primary database are also physically deleted. 3. Relationship limits: - The source object cannot be a child object (detail table object). Child object archiving depends on its parent object. - A standard object can have multiple archiving rules (one-to-many large objects). - Each large object can only be used by one standard object for archiving. 4. Field mapping limits: - Text, URL, location, and statistics fields from standard objects can be archived to text fields in large objects. - Person and number fields from standard objects must map to corresponding person and number fields in the large object. - Field mappings cannot be modified after saving. 5. Dependency cleanup order: To completely delete a large object, first delete or unbind its associated data archiving rule in the Admin Console.