The issues were caused by a broken server cluster configuration. The service is operating normally again.

We are currently working on a post mortem and will update this status accordingly.

Update:

The incident occurred between 16:00 and 16:46 CEST. Apps with a custom domain accessed using HTTPS were unavailable because the TLS certificate could not be served. The CORS Proxy was affected and was intermittently inaccessible as well. The Main Site was going up and down as well. Requests to apps were also failing around 17:50 as we tried to find out more about the incident. Storage was not affected.

We are sorry about this incident and are going to implement measures to prevent it to happen again.