At approximately 05:21:00 UTC on May 31, 2019, we were alerted that our upstream provider had experienced a fiber cut that impacted connectivity to our Toronto data center. At around 12:21:00 UTC, in an attempt to lessen the impact of the outage, a change to our route-policy was made which produced high CPU events on our edge routers in Dallas, Fremont, London, Newark, Singapore, and Tokyo 2 data centers.
Although we were able to revert the change immediately, the event did cause brief routing instability across our global backbone. The result was a momentary degradation of routing to the public internet.
We apologize for the connectivity issues experienced as a result of this change.