At 10:26pm UK local time, we started to receive automated reports of increased latency and packet loss to some destinations.
After some diagnostics, we discovered that traffic taking an inbound path via Cogent (AS174) to us was seeing increased latency of +900ms and 25-35% packet loss. Outbound data was unaffected.
We have disabled BGP announcements to Cogent to prevent inbound data travelling via this path. Latency and Packet Loss have dropped back to normal levels.
We will monitor and re-establish inbound communications once things stablise within Cogent.
Data not travelling inbound towards our network via Cogent was unaffected. All other paths remain stable.
Cogent suffered a power related issue in Slough which damaged some of their hardware. Overnight they dispatched engineers who replaced the damaged hardware and restored their service by 2:36am.
This affected all Cogent IP transit customers in the UK and was not restricted to Netnorth.
We continued to monitor throughout the night, and re-established our inbound paths at 10:55am.
No further disruption is anticipated.