Resolved. Between 19:42 and 22:04 UTC on October 13, the API could not reach its internal queue and cache service after a configuration change on one of our servers interrupted internal network traffic. Uploads and scheduled posts could not be processed during this window. The network was restored, the API was restarted, and all services have been operating normally since 22:04 UTC. A brief interruption (about 6 minutes, around 21:15 UTC) also affected the website, app and docs during a server restart. We have fixed the configuration that caused the issue, made the API reconnect automatically, and improved our monitoring so a similar problem is detected and fixed much faster. We apologize for the inconvenience.