Problem

Users are experiencing intermittent/inconsistent data loss while collaboratively editing drafts for pages and blog posts in Confluence Cloud, happening via web browsers and the mobile app. Drafts revert to a previous state, causing all subsequent updates and comments to disappear.

This issue is inconsistent and has not clear steps for reproducing.

Environment

• Confluence Cloud Drafts
  • Web Browsers
  • Confluence Mobile App

Steps to Reproduce

Note: below are generic steps, since the reported issue can't be constantly reproduced.

1. You can make changes to a page or blog post without saving them (leave them as a draft or a page with 'unpublished changes').
2. Later, wait for other collaborators to make changes to the content (either via web browser or mobile app).
3. After an extended period (several days or weeks), open the content via a web browser or mobile app, then refresh the page or restart the browser.
4. Observe that web changes were lost on the publish, or even before it.

Expected Results

• Changes made collaboratively should be saved accurately without reverting to previous states.
• Autosaved changes should appear in edit mode after refreshing or restarting the browser.

Actual Results

• Changes made to the draft are lost inconsistently after a few days or weeks while collaborating on the draft.
• Autosaved changes don't appear in edit mode initially, though they are visible under "... > View changes".

Workaround

Currently, users have two alternatives: 

1. Create a new page to collaborate on, considering the one with issues corrupted, and report the page to support for further investigation enablament.
2. Publish changes more consistently

Notes

• The issue has multiple reports but has not been consistently replicated by the support team or affected customers (in a web browser or mobile app).
• Force-crashing the browser or mobile app immediately after edits is the only known way to replicate the issue consistently within those drafts.

If you have additional questions or need support with this scenario, please reach out to our support team for a detailed investigation.