Issue Summary

Standard Jira Macro Repair (JMR) scans and updates published (current) page records in Confluence (bodycontent and ADF storage).

Collaborative editing drafts (content_status = 'draft') and the NCS (Node Collaboration Service) / Synchrony change history still hold the old Server ID.

When a user enters Edit mode, Confluence loads the cached draft rather than the clean published version. When published, this stale draft overwrites the database-repaired page, reintroducing the broken macro.

Steps to Reproduce

1. Create a Confluence page and add a Macro which contains a server ID
2. Edit the page and perform some changes
3. Do not publish the changes
4. Execute Jira Macro repair utility
5. The macros on current version of confluence pages will be fixed and APP ID would be updated to System JIRA
6. When you enter the edit mode on the updated page, Confluence loads the cached draft rather than the clean published version. When published, this stale draft overwrites the database-repaired page, reintroducing the broken macro.

Expected Results

The Macros on the Confluence Draft page should be picked up in scan and fixed during execution

Actual Results

When a user enters Edit mode, Confluence loads the cached draft rather than the clean published version. When published, this stale draft overwrites the database-repaired page, reintroducing the broken macro.

Workaround

Delete or publish the draft pages and run the Jira macro repair utility. However, it is a major challenge to do it at scale.