Stéphane BortzmeyerAnd just after, routing is removed as a possible cause (like DNS was before) and "We suspect that an inadvertent configuration change as the trigger event for this issue.'"
PEBKAC as usual?
Robert Edmonds@bortzmeyer There is a critical domain involved in the resolution of the CNAME chains used in Azure-related services, t-msedge.net. It has two nameserver IPs, both in the same /24. Hasn't been responding to me for a couple hours.
I think this is pretty fairly attributable as a "DNS failure" even if the underlying cause of the failure of the DNS service was a network failure.
@bortzmeyer Although, I guess in theory if DNS is a small part of the services that are affected by a network outage and those services wouldn't have worked anyway, had DNS been functional, it would be fair to say "not really a DNS outage, but a network outage".
Nicolas SAPA 
