I'm seeing a weird problem where traffic originating from within a Docker container located on a VM in the Google Cloud in europe-west1 and destined for a UK company based in Amazon's Cloudfront is incorrectly routing all the way out through the US to hit cloudfront in California, causing all kinds of unwanted latency and slowness.
$ ping destination.host.co.uk PING d3csmaahmfmvav.cloudfront.net (22.214.171.124) 56(84) bytes of data. 64 bytes from server-54-192-146-219.sfo4.r.cloudfront.net (126.96.36.199): icmp_seq=1 ttl=49 time=161 ms
HOWEVER, traffic originating from the host VM routes correctly to a European Cloudfront location:
$ ping destination.host.co.uk PING d3csmaafmvavjz.cloudfront.net (188.8.131.52) 56(84) bytes of data. 64 bytes from server-54-230-12-31.ams1.r.cloudfront.net (184.108.40.206): icmp_req=1 ttl=54 time=10.5 ms
I can't find anything in the network configuration for the Docker containers that would indicate why the containers would be sending traffic to the other side of the planet when the host itself is sending it where we would expect it to.
For the life of me I can't figure out why this is going on, hopefully someone can help point out what I'm missing.