The issue has been fixed and traffic levels have recovered.

After an administrative shutdown of an interface on one of the edge routers, every IP address got configured on every device, causing the traffic to be lost. We are still investigating the root cause for this behaviour and try to reproduce it in our lab.

The issue has been identified to one edge router. We have isolated the affected router and investigate why the traffic was still being forwarded towards the broken router.

We are investigating a routing outage in CGN1. Fabric traffic is not affected, but all connections towards the internet may show packet loss or is being dropped at onr of the edge routers.