---
title: "Keep Created Date When Migrating: System Fields & Import APIs"
slug: keep-created-date-when-migrating-system-fields-import-apis
date: 2026-11-17
author: Rishabh Makhar
categories: [Migration Guide, Checklist, From The Migration Trenches]
excerpt: "Standard create APIs stamp every migrated record with today's date. Learn the import path per platform that preserves original timestamps, authors and history."
tldr: "Created date and author are system fields. To keep original history, use each platform's sanctioned import path, not a generic create call, whether a script, an AI agent or an MCP tool makes it."
canonical: https://clonepartner.com/blog/keep-created-date-when-migrating-system-fields-import-apis
---

# Keep Created Date When Migrating: System Fields & Import APIs


# Keep Created Date When Migrating: System Fields & Import APIs

Every record you create through a standard API endpoint, or let an [AI agent create through an MCP tool](https://clonepartner.com/blog/blog/ai-agent-data-migration-via-mcp-the-2026-guide), lands in the destination stamped with today's date and attributed to whichever account made the call. A five-year-old support ticket becomes brand new. The original author disappears. SLA reports, ageing views and first-contact metrics break quietly.

To keep the original created date when migrating, you have to use the platform's import path or audit permission that allows writes to system fields. Each platform has one specific path: sometimes a separate API, sometimes a special permission, sometimes a CSV importer. Miss it, and fixing the dates later usually means deleting and re-loading.

This article maps that path for six platforms, explains what breaks when you skip it, and ends with a pre-migration checklist for preserving timestamps and authorship.

*Last verified: 17 November 2026. MCP server capabilities, API token policies and field writability change. Check the linked vendor documentation before you migrate.*

## What are system fields, and why can't you just set them?

**System fields** are platform-owned metadata, such as created date, created by, modified date and modified by, that the platform fills in when a record is inserted or changed. Salesforce describes them as read-only fields that "are automatically updated during API operations" ([developer.salesforce.com](https://developer.salesforce.com/docs/platform/object-reference/guide/system-fields.html)). Zendesk marks a ticket's `created_at` as read-only ([developer.zendesk.com](https://developer.zendesk.com/api-reference/ticketing/tickets/tickets/)).

So a ticket created through Zendesk's normal endpoint is stamped with the current time. That is correct for genuinely new records. For a migration, it re-dates your entire history to migration day.

```bash
# Zendesk: standard ticket endpoint. created_at is read-only here,
# so the new ticket holds the current time.
POST /api/v2/tickets
{
  "ticket": {
    "subject": "Network outage",
    "created_at": "2019-03-15T09:00:00Z"
  }
}

# Zendesk: Ticket Import API (admins only). created_at can be set.
POST /api/v2/imports/tickets
{
  "ticket": {
    "subject": "Network outage",
    "created_at": "2019-03-15T09:00:00Z"
  }
}
```

This is why a migration script can look successful while it loses history. It creates records, gets success responses, and copies bodies and custom fields correctly. What it does not keep is the original timeline.

A general-purpose AI agent meets the same boundary. [Vendor MCP servers are designed for conversational work on a few records at a time](https://clonepartner.com/blog/blog/mcp-server-limits-that-break-data-migrations-2026), and their create tools call the normal create path, not the import path. Intercom's MCP server has no conversation-create tool at all, and articles created through it are attributed to the admin who authorized the connection ([developers.intercom.com](https://developers.intercom.com/docs/guides/mcp)).

The standard path has a second problem. Records created through it behave like new activity: triggers fire, timers start, and customers can be emailed about issues resolved years ago.

## How each platform preserves (or blocks) original dates and authors

The table maps the system field, the path that can set it, and the constraints. Detail follows.

| Platform | Date field(s) | Author field(s) | History-preserving path | Key constraints |
|---|---|---|---|---|
| **Zendesk** | `created_at`, `updated_at`, `solved_at`, comment `created_at` | Requester, and comment `author_id` in the import examples | Ticket Import API (`/api/v2/imports/tickets`), admins only | Ticket dates before 1970 rounded up to 1970. Comment dates cannot be before 1970 or in the future. 100 tickets per bulk request. Metrics and SLAs not supported on imported tickets. |
| **HubSpot** | `Create date` | — | Writable only on **deals** | Read-only on every other object. Use a custom property for the original date. |
| **Salesforce** | `CreatedDate`, `LastModifiedDate` | `CreatedById`, `LastModifiedById` | Insert with "Set Audit Fields upon Record Creation" enabled and granted | Insert only. `SystemModstamp` can never be set. Earliest valid audit date is 1 January 1970 GMT. |
| **Jira Cloud** | Created, mapped to Date created | — | CSV import through External System Import | Date format entered in the wizard must match the CSV, or the created date defaults to the current date. |
| **Microsoft Dataverse** | `createdon` via `overriddencreatedon` | `createdby` cannot be imported | Import mapped to `overriddencreatedon`, with the `prvOverrideCreatedOnCreatedBy` privilege | `modifiedon`, `createdby` and `modifiedby` cannot be imported. Use custom columns. |
| **Intercom** | Conversation `created_at` | MCP-created articles fixed to the authorizing admin | REST Create Conversation with a `created_at` UTC Unix timestamp | MCP server has no conversation-create tool. |

### How does Zendesk's Ticket Import API preserve created dates?

Zendesk's Ticket Import API has two endpoints: `POST /api/v2/imports/tickets` and `POST /api/v2/imports/tickets/create_many`. Both are allowed for admins only. You can set `solved_at`, `updated_at` and `created_at` on tickets, and `created_at` on comments ([developer.zendesk.com](https://developer.zendesk.com/api-reference/ticketing/tickets/ticket_import/)).

The constraints, from the same page:

- A ticket `created_at` or `updated_at` before 1970 is rounded up to 1970.
- A comment `created_at` cannot be before 1970 or in the future.
- "If a ticket with a status other than closed is imported, triggers won't run on the imported ticket. However, if the ticket is updated after import, triggers will resume."
- No notifications are sent to users CC'd on imported tickets.
- The bulk endpoint accepts up to 100 tickets per request.

Requests count against your plan's Support API rate limit: 200 requests per minute on Team, 400 on Growth and Professional, 700 on Enterprise, and 2,500 on Enterprise Plus or with the High Volume API add-on ([developer.zendesk.com](https://developer.zendesk.com/api-reference/introduction/rate-limits/)).

The most important reporting constraint: "Zendesk metrics and SLAs are not supported for imported tickets." Running them on imported tickets gives "incomplete and inaccurate data". Zendesk recommends tagging imported tickets, for example `migrated_historical`, and excluding them from reports ([developer.zendesk.com](https://developer.zendesk.com/api-reference/ticketing/tickets/ticket_import/)).

> [!WARNING]
> **Zendesk API token sunset:** Accounts created on or after 28 July 2026 cannot create or use API tokens. Since 27 October 2026, no account can create new tokens. All remaining tokens are deactivated permanently on 30 April 2027, and OAuth replaces them ([support.zendesk.com](https://support.zendesk.com/hc/en-us/articles/10851263566234)). Any migration script that authenticates with an API token needs to move to OAuth before then.

### Can you set HubSpot's Create date during a migration?

HubSpot says: "The Create date property is read-only for all objects except deals" ([knowledge.hubspot.com](https://knowledge.hubspot.com/records/update-a-property-value-for-a-record)). For contacts, companies, tickets and other objects, the original date cannot go into the native field.

The standard workaround:

1. Create a **custom date property**, such as `original_created_date`.
2. Populate it with the legacy creation date during import.
3. Use that property in reports and filters instead of the system `Create date`.

This fixes reporting you control. Views and automations built on the native `Create date` still see the import date.

Two more HubSpot caveats matter for the load. Companies created through the API "will not be deduplicated by the Company domain name property", and import rows without an email or Record ID become new records ([knowledge.hubspot.com](https://knowledge.hubspot.com/records/deduplication-of-records)). Associations need the IDs of both records, so parents must exist before children can be linked ([developers.hubspot.com](https://developers.hubspot.com/docs/api-reference/latest/crm/associations/associate-records/guide)).

### How do you set Salesforce audit fields during a data load?

Salesforce's audit fields, `CreatedDate`, `CreatedById`, `LastModifiedDate` and `LastModifiedById`, can be set on insert. The setup ([developer.salesforce.com](https://developer.salesforce.com/docs/platform/object-reference/guide/system-fields.html)):

1. From Setup, enter **User Interface** in the Quick Find box and select User Interface.
2. Enable the "Set Audit Fields upon Record Creation" and "Update Records with Inactive Owners" user permissions.
3. In the permission set or profile that will load the data, enable **Set Audit Fields upon Record Creation**.
4. Create records through the API and set their audit fields.

`SystemModstamp` is the one audit field you cannot set. Salesforce lists the objects that accept audit values, including Account, Case, CaseComment, Contact, Lead, Opportunity, Task, Event and custom objects. Audit date fields have an earliest valid date of 1970-01-01T00:00:00Z GMT ([developer.salesforce.com](https://developer.salesforce.com/docs/platform/object-reference/guide/system-fields.html)). Keeping records owned by inactive users needs "Update Records with Inactive Owners".

The practical sequence for a load:

1. Enable and grant the audit-field permission as above.
2. Set validation rules, lead and case assignment rules and territory assignment rules to inactive, and bypass triggers with a custom setting checkbox, as Salesforce's large-data-volume module describes ([trailhead.salesforce.com](https://trailhead.salesforce.com/content/learn/modules/large-data-volumes/load-your-data)).
3. Insert records with their audit values. Use upsert by external ID so a failed batch can be retried without duplicates, and so child records can reference parents by external ID ([developer.salesforce.com](https://developer.salesforce.com/docs/platform/api-rest/guide/dome-upsert.html)).
4. Turn validation, workflow and assignment rules back on, and reset the trigger checkbox.
5. Remove the audit-field permission from the loading user.

Because audit fields are set on insert only, a record loaded without step 1 keeps today's date and the loading user as creator. The fix is to delete and re-insert it.

### How does Jira Cloud keep original created dates on imported issues?

Atlassian's documented way to keep an original created date is a CSV import through **External System Import**. On the project step, enter a date format that matches your CSV, then map the Created column to the **Date created** field ([support.atlassian.com](https://support.atlassian.com/jira/kb/how-to-retain-created-date-while-importing-issues/)).

The trap is in the format. Atlassian says the created date defaults to the current date if you don't map the field, or if the CSV's date format is incorrect. The CSV format may also differ from the default shown in the UI. Test with a few rows before the full file.

### How does Microsoft Dataverse preserve creation dates on import?

Dataverse keeps the original creation date through a separate column, **`overriddencreatedon`**. Map your source creation date to it, and during import the record's `createdon` is set to that value, while `overriddencreatedon` records when the import ran. If nothing is mapped to it, `createdon` is the import time ([learn.microsoft.com](https://learn.microsoft.com/en-us/power-apps/developer/data-platform/run-data-import)).

Overriding `createdon` needs the **`prvOverrideCreatedOnCreatedBy`** privilege. Despite the name, overriding `createdby` is not currently supported. You cannot import data into `modifiedon`, `createdby` or `modifiedby`. Microsoft's guidance is to create custom columns for who created and modified the record, and when it was modified ([learn.microsoft.com](https://learn.microsoft.com/en-us/power-apps/developer/data-platform/run-data-import)).

### Can Intercom's MCP server migrate conversations with original dates?

Intercom's REST API can. The Create Conversation endpoint takes `created_at` as a UTC Unix timestamp. Intercom recommends the field only for migrating past conversations from another source, and uses the current time when it is missing ([developers.intercom.com](https://developers.intercom.com/docs/references/rest-api/api.intercom.io/conversations/createconversation)).

The MCP server cannot. Its write tools add internal notes and create or update Help Center articles, and there is no tool to create conversations or contacts ([developers.intercom.com](https://developers.intercom.com/docs/guides/mcp)). The [help desk guide](https://clonepartner.com/blog/blog/can-an-ai-agent-migrate-your-help-desk-zendesk-intercom-freshdesk) covers Intercom, Zendesk and Freshdesk in more depth.

The gap in a sentence: the REST API supports migration, and the MCP server was not built for it.

## What is author collapse?

**Author collapse** is when every migrated record is attributed to the single account that made the API calls, instead of the original authors. It is separate from timestamp loss. A record can have the right date and the wrong author.

Migrate 10,000 comments written by 15 agents through one connection without author mapping, and the destination shows one person writing all 10,000.

How it shows up by platform:

- **Zendesk:** the Ticket Import API's examples include an `author_id` on each comment ([developer.zendesk.com](https://developer.zendesk.com/api-reference/ticketing/tickets/ticket_import/)). Map source authors to Zendesk user IDs before the load, and confirm authorship in the sample.
- **Intercom:** articles created through MCP are attributed to the admin who authorized the connection, and the author cannot be set per call ([developers.intercom.com](https://developers.intercom.com/docs/guides/mcp)).
- **Dataverse:** `createdby` and `modifiedby` cannot be imported. Store legacy authorship in custom columns ([learn.microsoft.com](https://learn.microsoft.com/en-us/power-apps/developer/data-platform/run-data-import)).

The effects are operational, not cosmetic:

- **Agent performance reports** credit thousands of historical interactions to one user.
- **"Replied by" labels** show the wrong name in threads and portals.
- **Audit trails** lose who said what, which is a compliance problem in regulated industries.

Preserving authorship means using the import path, creating all users in the destination first, and mapping source user IDs to destination user IDs before the load. Where the author cannot be set, a fallback is to prefix the comment body with the original author and time, for example `[Originally by: Jane Smith, 2019-03-15]`, or to store it in a custom field such as **Legacy Author**.

> [!TIP]
> A Legacy Author field or prefix keeps the information readable. It is not native audit history: filters, permissions, @mentions and system reports still use the platform's real author fields.

## A taxonomy of migration history failures

Naming the failure modes helps you decide which checks to run after the load. There are five:

1. **Timestamp displacement.** Records carry the import date instead of the original. Cause: the standard create endpoint. Check: compare the oldest record date in source and destination.
2. **Author collapse.** Records are attributed to the migration user. Cause: one credential without author mapping. Check: count records created by the migration user and ask whether the volume is plausible.
3. **Trigger and automation pollution.** History activates business logic built for new records. Cause: standard endpoints, or later updates to imported records. Check: watch outbound email and automation logs during the run.
4. **Broken attachment references.** Inline images and attachments point at expired source URLs. Cause: copying URLs instead of files. Check: request every attachment URL in the destination after import.
5. **Missing archived records.** Records exist in the source but are never extracted. Cause: the wrong export endpoint. Zendesk's List Tickets does not include archived tickets, so use the Incremental Ticket Export ([developer.zendesk.com](https://developer.zendesk.com/api-reference/ticketing/tickets/tickets/)). Check: compare record counts per object type.

## What breaks when every record says "created today"?

Re-dated history corrupts operational data in ways that are hard to spot and expensive to fix.

**SLA and metrics reports.** A ticket "created" today and resolved five minutes later looks like a five-minute resolution, not the three days it took. With the right import path, dates are kept, but Zendesk still says metrics and SLAs are not supported for imported tickets ([developer.zendesk.com](https://developer.zendesk.com/api-reference/ticketing/tickets/ticket_import/)). Either way, history needs a tag and an exclusion.

**Ageing and backlog views.** Any view sorted by created date, such as "open issues by age", shows no old records. The historical backlog looks brand new, and teams cannot tell new work from migrated history.

**First-response metrics.** Time-to-first-reply is the gap between creation and the first agent comment. If both carry migration-day timestamps, the history becomes useless for benchmarking or capacity planning.

**Triggers on historical data.** On Zendesk's standard endpoint, business rules apply, "which can include sending email notifications to your end users" ([developer.zendesk.com](https://developer.zendesk.com/api-reference/ticketing/tickets/tickets/)). An email about a three-year-old issue cannot be recalled by deleting the ticket.

**Reporting windows.** If dashboards filter to "last 90 days", the whole migrated volume lands in the current window and skews averages and trends.

## Timezone normalization: a specific risk

Timezone handling is a common silent error in migrations that otherwise use the right import path. A timestamp exported without an offset is ambiguous, and the destination will read it in whatever zone its importer assumes.

Two documented examples:

- **Intercom** takes `created_at` as a UTC Unix timestamp ([developers.intercom.com](https://developers.intercom.com/docs/references/rest-api/api.intercom.io/conversations/createconversation)). A local time converted without its offset shifts every conversation.
- **Salesforce** audit date fields have an earliest valid date of 1970-01-01T00:00:00Z GMT, and Salesforce notes that these values are offset by your time zone ([developer.salesforce.com](https://developer.salesforce.com/docs/platform/object-reference/guide/system-fields.html)). Records near that boundary need care.

The practice: confirm the timezone of every source export, normalize timestamps to UTC with an explicit offset before the load, and spot-check three to five records whose original local time you know.

## Upsert and idempotency patterns across platforms

A retry should update, not duplicate. Each platform offers a different mechanism, and some offer less than you might assume.

| Platform | Mechanism | Notes |
|---|---|---|
| **Zendesk** | `Idempotency-Key` header on ticket creation | "Keys expire after two hours." A duplicate key sent later creates a new ticket ([developer.zendesk.com](https://developer.zendesk.com/api-reference/ticketing/introduction/)). Ticket `external_id` values "don't have to be unique", so they are not a dedupe key ([developer.zendesk.com](https://developer.zendesk.com/api-reference/ticketing/tickets/tickets/)). |
| **HubSpot** | Email for contacts, Record ID on import | Companies created through the API are not deduplicated by domain. Import rows without an email or Record ID become new records ([knowledge.hubspot.com](https://knowledge.hubspot.com/records/deduplication-of-records)). |
| **Salesforce** | Upsert by external ID | Makes a load idempotent and lets children reference parents by external ID ([developer.salesforce.com](https://developer.salesforce.com/docs/platform/api-rest/guide/dome-upsert.html)). |
| **Dataverse** | Upsert message on the Web API | Listed as a table operation alongside Create and Update ([learn.microsoft.com](https://learn.microsoft.com/en-us/power-apps/developer/data-platform/reference/entities/subject)). |

Where the platform gives you nothing durable, the loader has to supply it. Store the destination record ID against the source record ID in a mapping table that persists outside any session, check it before every write, and load parents before children.

## Handling attachments and expiring URLs

A preserved timestamp is little use if the attachments are broken. Many sources protect files with expiring URLs.

Notion-hosted file URLs expire after one hour ([developers.notion.com](https://developers.notion.com/reference/file-object)). Zendesk attachment uploads return a token that must be attached to a ticket within 60 minutes ([developer.zendesk.com](https://developer.zendesk.com/api-reference/ticketing/tickets/ticket-attachments/)). A queue that takes 90 minutes to reach a record will miss both windows.

A script that copies file URLs instead of the files ships links that break. The correct pipeline downloads the file, uploads it to the destination, takes the new URL, rewrites the record body, and only then writes the record.

## Pre-migration checklist for preserving history

Work through this list before writing migration code. Each item maps to a failure class above.

- [ ] **Identify system fields on the destination.** Find the exact names for created date, created by, modified date, modified by and resolved date, and which are writable.
- [ ] **Find the history-preserving import path.** A separate import API (Zendesk Ticket Import API), a permission (Salesforce audit fields), an import tool (Jira External System Import) or a dedicated column (Dataverse `overriddencreatedon`). If you cannot name it for an object, assume native history will not survive.
- [ ] **Check permissions.** Zendesk's Ticket Import API is admin-only. Salesforce needs "Set Audit Fields upon Record Creation". Dataverse needs `prvOverrideCreatedOnCreatedBy`.
- [ ] **Map source users to destination users.** Create every agent, admin and end user in the destination first, including inactive ones, and keep the ID map.
- [ ] **Normalize timestamps.** Convert to UTC with explicit offsets, and spot-check five records whose local time you know.
- [ ] **Confirm date formats.** Jira's import needs the format entered in the wizard to match the CSV. Zendesk rounds pre-1970 ticket dates up to 1970. Intercom expects a UTC Unix timestamp.
- [ ] **Decide what happens to fields that cannot be preserved.** Dataverse cannot import `modifiedon`, `createdby` or `modifiedby`. HubSpot's Create date is read-only except on deals. Plan custom fields now and tell stakeholders before go-live.
- [ ] **Tag imported records.** Zendesk recommends tagging imported tickets and excluding them from reports. Do the same on every platform.
- [ ] **Disable triggers, automations and notifications.** Salesforce's large-data-volume guidance is to set validation and assignment rules inactive and bypass triggers during the load. On Zendesk, triggers won't run on imported tickets whose status is anything other than closed, and resume if the ticket is updated after import.
- [ ] **Validate attachment handling.** Download and re-upload files rather than passing URLs. Notion URLs expire in one hour. Zendesk upload tokens must be attached within 60 minutes.
- [ ] **Prove extraction completeness.** Use Zendesk's Incremental Ticket Export to include archived tickets, and compare record counts per object type.
- [ ] **Build a source-to-destination ID map.** Required for idempotent retries and for loading children after parents.
- [ ] **Run a sample and check dates and authors.** Check the oldest open item, oldest closed item, a reopened item and an attachment-heavy item. These surface timestamp and author bugs fastest.

For field mapping, see our [data migration mapping cheat sheet](https://clonepartner.com/blog/blog/data-migration-mapping-cheat-sheet-sample-scripts). For post-migration validation, use our [QA checklist](https://clonepartner.com/blog/blog/help-desk-data-migration-qa-checklist). If thread fidelity is in scope, pair this with our guide to [preserving email history, threads and activity logs](https://clonepartner.com/blog/blog/preserving-email-history-threads-activity-logs-in-a-crm-migration).

## Where AI genuinely helps

AI is useful on this problem, in the right role. It can read a destination's documentation and schema and draft the list of system fields and the import path for each object. It can write the transformation code that normalizes timestamps and maps user IDs, plus the tests for it, and that code then runs outside the model. It can also review a sample and flag records whose dates or authors look wrong.

What it should not do is make the writes itself through tools built for conversational work. The pattern that works: AI drafts, a person reviews, and deterministic code runs the job.

## How ClonePartner preserves data history

ClonePartner's migration engine already knows the import paths that preserve history, such as Zendesk's Ticket Import API, and the permissions that let original dates and authors be set, such as Salesforce's "Set Audit Fields upon Record Creation".

It also handles the details around them: parent-before-child load order, a source-to-destination ID map that persists outside any session, idempotent loads by external ID, attachment handling within upload-token and signed-URL expiry windows, pausing and restoring destination automations, and each platform's rate limits.

Before the full migration, you get **unlimited sample migrations**, included in the price, to check dates, authors and statuses in the destination. The field mapping is frozen once you approve the sample, and the same mapping runs the full load and the delta. Validation includes record-count parity, field-level checks, referential integrity and a reconciliation report.

If timestamps or authorship look wrong in a sample, the engineer fixes the mapping and runs it again. That is the point of a sample: catching these problems before production.

> [!NOTE]
> **Two ways to run it.** If your source and destination are a pair ClonePartner Express supports, you can run the migration yourself: [Gorgias](https://gorgias-migration.clonepartner.com), [Front](https://front-migration.clonepartner.com), [Superhuman Docs](https://superhuman.clonepartner.com) or [Google Docs](https://google.clonepartner.com), each with a free sample on your real data first. For any other pair, or custom fields, relational records, permissions or compliance requirements, [talk to a migration engineer](https://clonepartner.com/talk-to-us?duration=30&utm_source=blog&utm_medium=button&utm_campaign=demo_bookings&utm_content=cta_click&utm_term=demo_button_click) about a fixed-price managed migration.

> Book a free 30-minute scoping call. We'll map the history-preserving import path for your source and destination and give you a fixed-price proposal, with no hourly billing.
>
> [Talk to us](https://clonepartner.com/talk-to-us?duration=30&utm_source=blog&utm_medium=button&utm_campaign=demo_bookings&utm_content=cta_click&utm_term=demo_button_click)

## Frequently asked questions

### How do I keep the original created date when migrating Zendesk tickets?

Use Zendesk's Ticket Import API (POST /api/v2/imports/tickets), not the standard ticket endpoint. It is admin-only and lets you set created_at, updated_at and solved_at on tickets, and created_at on comments. On the standard endpoint, created_at is read-only, so tickets get the current time. Zendesk says metrics and SLAs are not supported for imported tickets, and recommends tagging them and excluding them from reports.

### Can you set Salesforce CreatedDate during a data migration?

Yes, but only on insert. From Setup, open User Interface and enable the "Set Audit Fields upon Record Creation" user permission, then grant it in the permission set or profile that loads the data. SystemModstamp can never be set. Because audit fields are set only on insert, a record loaded without this setup keeps today's date unless you delete and re-insert it.

### Is HubSpot's Create date writable via API?

Only for deals. HubSpot says the Create date property is read-only for all objects except deals. For contacts, companies, tickets and other objects, create a custom date property such as Original Created Date, populate it during import, and use it in reports instead of the native Create date.

### How do you retain the created date when importing issues into Jira Cloud?

Atlassian's documented way is a CSV import through External System Import. Enter a date format that matches your CSV on the project step, and map the Created column to the Date created field. If the field is not mapped or the format is wrong, the created date defaults to the current date.

### What is author collapse in a data migration?

Author collapse is when every migrated article, comment or record is attributed to the single account that made the API calls, instead of the original author. To prevent it, use import paths that take author IDs, create users in the destination first, and map source user IDs to destination user IDs before the load.
