We have today received an RFO from our upstream transit provider. The cause of this incident was an upstream international transit outage between AS9790 and AS4826 caused by AS9790 bringing a new BGP peer via AS6939 online.
Upon AS9790 bringing that new peer with AS6939 online, additional BGP routes got re-advertised by AS9790 to AS4826 resulting in AS9790 prefix limits with AS4826 being reached and accordingly BGP sessions automatically shut down. The impact of this was the loss of about 18% of international transit.
As soon as the prefix limits were noticed the change was rolled back to restore connectivity and a post-incident review has been conducted to identify and address opportunities for improvement.