We've had reports of customers having problems routing to some IPs within Amazon
We're altering our routing
Disabling our direct link to Amazon has re-routed the traffic to our transit links, and this is looking stable. We saw intermittent problems with routes to Amazon stopping at the Amazon router, this looked like a LAG issue within Amazon as traffic was working fine from some devices whilst not from others. We'll get in touch with Amazon.
We have reached out to Amazon, but they seem rather busy with other things... As our routing to/from them is working via transit connection this is no longer affecting services.
Amazon have advised us that they have fixed the cause the problems we were seeing. We will re-enable this interlink at around 10AM Tuesday 4th November.
We had planned to bring up the BGP session, but the Amazon router isn't responding, so we'll investigate further.
Separate to the routing problem, both of our Amazon interconnects are being migrated to different equipment at Amazon's end. They are out of action for the moment. We still have peering to Amazon over LINX.