The incident is resolved. It was caused by a member's maintenance resulting in a loop. This should have been caught by our switches' filters. Unfortunately a bug in handling neighbor discovery caused high CPU usage on the switches serving the respective customer. This in turn ...

The loop has been mitigated.
It remains unclear why implemented protection mechanisms did not prevent the loop from causing issues.

We have fixed the issue and are monitoring the affected services.

Some customer seems to have formed a loop.

We are currently investigating an issue reported by our monitoring systems.