This morning we experienced that our DanDomain App wasn't working properly. Everything is back to normal now.
The Clerk.io service was running normally so this issue only affected customers on DanDomain running our native DanDomain App.
DanDomain requires App providers to build an intermediate service to bridge DanDomain to external services. This service was hosted at Linode and not AWS in our std deployment environment. During a scheduled maintenance operation performed by the Linode team to patch the Spectre and Meltdown vulnerabilities, the servers IP configuration was corrupted and the server thus didn't recognize it's IPv4 address.
Due to the complexity of the error, it took some time for the Linode team to discover and resolve the problem.
We are working on moving all our external Apps from Linode to AWS and monitor them as part of our own infrastructure. This is not a type of error that can apply to other platforms than DanDomain.