Updated 23:04 UTC: All portals are back online. The interruption was caused by a faulty load-balancing component that affected just over half of all customer portals. We were able to replace the component within 3 minutes of failure, but unfortunately it takes a few minutes for network traffic to be re-routed. Apologies for the disruption. We are already working on further redundancy on our load-balancing layer.
We're working on elevated network-related errors, and will be back to normal shortly. Apologies for the interruption.