The change occurred as planned at 15:00 UTC. However, at the same time
the CloudAMQP round-robin DNS entry was erroneously removed, causing
pulse.mozilla.org, which is aliased to that entry, to fail to resolve.
This caused TaskCluster to fail, which in turn caused a brief tree
closure. The DNS entry was restored around 15:19, and TaskCluster was
back up around 15:26. Everything appears to be functioning correctly,
and, as mentioned, this kind of problem should now be in the past.
Mark