Hi Guys,
Several networks peered with Amazon in London have noticed sporadic but easily replicated slowness (a few KB/s) on around 25% of traffic from S3.
This has been going on for over a week; email to your NOC a couple of days ago hasn't received a response but issue looks quite widespread.
Happy to share more details, but a simple wget loop to an S3 resource in eu-west-* will fail between 10-30% of the time. I've had this confirmed by at least 5 other ASNs in the region.
We don't see the same problem in our other peering locations, and we see this issue on more than one London IX.
Regards,