On Wednesday, October 22, 2025, between ~12:50–13:30 ET, some customers in the US and EU regions experienced elevated errors and timeouts from the Photo Center API following a production upgrade. This appeared in the UI as broken upload modals and images not loading in Photo Center. Image delivery to end users was not affected. Service returned to normal after the upgrade was rolled back.
A platform upgrade changed application behavior and led to database connection pool exhaustion in the Photo Center API, causing requests to queue and time out. Rolling back the deployment in affected regions—and then across remaining regions as a precaution—restored normal performance by ~13:30 ET.
All times ET + 24 hour clock
| Time | Event |
|---|---|
| 12:45 | Upgrade deployment started in US and EU regions. |
| 12:50 | Alerts and customer reports of API errors/timeouts begin. |
| 13:00 | Investigation underway |
| 13:18 | Rollback to the prior version initiated in affected regions. |
| 13:28 | Rollback completed; service stabilized and alerts cleared. |
| 13:30 | Proactive rollback initiated in remaining regions; incident resolved |