On August 27, 2025, at 12:31 UTC, our Support team received reports from customers that they were encountering issues when attempting to access their models. During the incident, new models were unable to load assets. Access to the platform via UX apps was unaffected. The issue was global, affecting all regions.
Our investigation revealed that an internal change had been made to our Cloudflare configuration shortly before the first case was received. As a result of this change, requests relating to content delivery were receiving an error as they were unable to resolve.
We began rolling back the change at 13:10 UTC. The rollback was expected to take around 20 minutes, however, due to the complexity of the change and the related rollback steps, it took longer than expected. The rollback was completed at 14:09 UTC.
After completing the rollback, we confirmed that the underlying issue was resolved. However, we found that a proportion of the underlying infrastructure had cached the error. Initially, we manually moved workspaces from the affected infrastructure. However, at 15:36 UTC, we made the decision to move all workspaces in the regions below to fresh infrastructure.
eu4: Cloud - Europe
us5: Cloud - US East
ap1: Cloud - Japan
us7: Cloud - US
All migrations were completed at 21:44 UTC.
Following the incident, we've taken immediate action to address the root cause and are implementing further improvements to strengthen our platform. A key difference between our testing and production environments prevented this issue from being caught during testing. We have now corrected this discrepancy, ensuring our testing is more comprehensive going forward.
In addition to this, we are rolling out further improvements:
We are deploying a fix that stops temporary delivery errors from being cached, allowing them to resolve automatically.
We are enhancing our monitoring to catch and resolve potential issues faster.
We sincerely apologize for the impact this incident has had on your business operations, and we hold ourselves accountable for ensuring it doesn't happen again. We deeply appreciate your partnership and patience. If you have further questions, our team is ready to help via the Anaplan Support portal.