I've noticed in the past few days downloads from Cloudfront (Amazon's CDN) have become very slow - testing from two different Bigpipe Gigabit UFB connections, both in Auckland. I haven't had speed issues with Cloudfront before now, it's always been very fast.

An example file you can download is https://download.docker.com/linux/ubuntu/dists/xenial/pool/stable/amd64/docker-ce_17.06.2~ce-0~ubuntu_amd64.deb - I've been seeing speeds of ~20KB/s on that download, whereas cloudfront previously would get close to line speed. I'm also seeing slow performance on other files hosted on cloudfront, for example https://web.kamihq.com/web/notablecompiled/2.0.8993-20170908152259/notable.js , or you can do a online test at https://cloudharmony.com/speedtest-for-aws:cloudfront 

Here's what one of these slow downloads looks like:

wget "https://download.docker.com/linux/ubuntu/dists/xenial/pool/stable/amd64/docker-ce_17.06.2~ce-0~ubuntu_amd64.deb"
--2017-09-10 13:06:26-- https://download.docker.com/linux/ubuntu/dists/xenial/pool/stable/amd64/docker-ce_17.06.2~ce-0~ubuntu_amd64.deb
Resolving download.docker.com (download.docker.com)... 52.85.40.20, 52.85.40.134, 52.85.40.17, ...
Connecting to download.docker.com (download.docker.com)|52.85.40.20|:443... connected.
HTTP request sent, awaiting response... 200 OK
Length: 20487416 (20M) [application/x-debian-package]
docker-ce_17.06.2~ce-0~ubuntu_amd64.deb.1 17%[== ] 3.33M 5.99KB/s eta 16m 14s

And a traceroute:

~ $ traceroute download.docker.com
traceroute: Warning: download.docker.com has multiple addresses; using 52.85.40.17
traceroute to d2h67oheeuigaw.cloudfront.net (52.85.40.17), 64 hops max, 52 byte packets
1 router.asus.com (192.168.5.1) 1.421 ms 1.126 ms 0.959 ms
2 219.88.232.254 (219.88.232.254) 2.258 ms 2.494 ms 2.087 ms
3 122.56.60.70 (122.56.60.70) 2.439 ms 2.319 ms 2.113 ms
4 122.56.60.71 (122.56.60.71) 3.005 ms 3.094 ms 3.556 ms
5 ae11-201.tkbr11.global-gateway.net.nz (122.56.118.149) 2.601 ms 2.297 ms 1.990 ms
6 122.56.119.109 (122.56.119.109) 2.414 ms 2.321 ms 2.113 ms
7 xe0-0-8.sebr3.global-gateway.net.nz (122.56.119.126) 29.939 ms
xe0-0-0.sebr3.global-gateway.net.nz (202.50.232.194) 25.698 ms
xe0-0-1.sebr3.global-gateway.net.nz (202.50.232.14) 25.278 ms
8 122.56.119.86 (122.56.119.86) 26.066 ms 25.887 ms 25.301 ms
9 * i-0-4-0-0.sydp01.bi.telstraglobal.net (202.84.223.10) 57.872 ms
i-0-11-0-0.sydp01.bi.telstraglobal.net (202.84.220.210) 57.471 ms
10 i-0-1-0-7.sydp-core03.bi.telstraglobal.net (202.84.223.9) 58.721 ms 57.695 ms 58.506 ms
11 * i-0-1-0-49.sydo-core04.bi.telstraglobal.net (202.84.223.22) 58.588 ms 56.601 ms
12 unknown.telstraglobal.net (134.159.125.177) 57.903 ms 57.328 ms 56.889 ms
13 unknown.telstraglobal.net (134.159.125.178) 57.803 ms * 56.671 ms
14 52.95.37.96 (52.95.37.96) 63.698 ms
52.95.37.112 (52.95.37.112) 59.053 ms
52.95.37.96 (52.95.37.96) 61.355 ms
15 52.95.37.69 (52.95.37.69) 55.969 ms * *
16 54.240.204.215 (54.240.204.215) 67.746 ms 69.079 ms 67.718 ms
17 * * *
18 * * *
19 * * *
20 * server-52-85-40-17.mel50.r.cloudfront.net (52.85.40.17) 68.350 ms 67.546 ms

If I ping, I'm seeing ~30% packet loss:

~ $ ping download.docker.com
PING d2h67oheeuigaw.cloudfront.net (52.85.40.44): 56 data bytes
Request timeout for icmp_seq 0
64 bytes from 52.85.40.44: icmp_seq=1 ttl=239 time=69.138 ms
64 bytes from 52.85.40.44: icmp_seq=2 ttl=239 time=66.691 ms
64 bytes from 52.85.40.44: icmp_seq=3 ttl=239 time=67.848 ms
Request timeout for icmp_seq 4
64 bytes from 52.85.40.44: icmp_seq=5 ttl=239 time=67.675 ms
64 bytes from 52.85.40.44: icmp_seq=6 ttl=239 time=67.487 ms
64 bytes from 52.85.40.44: icmp_seq=7 ttl=239 time=67.434 ms
64 bytes from 52.85.40.44: icmp_seq=8 ttl=239 time=67.618 ms
64 bytes from 52.85.40.44: icmp_seq=9 ttl=239 time=67.117 ms
64 bytes from 52.85.40.44: icmp_seq=10 ttl=239 time=67.248 ms
64 bytes from 52.85.40.44: icmp_seq=11 ttl=239 time=68.288 ms
64 bytes from 52.85.40.44: icmp_seq=12 ttl=239 time=67.206 ms
Request timeout for icmp_seq 13
Request timeout for icmp_seq 14
Request timeout for icmp_seq 15
Request timeout for icmp_seq 16
Request timeout for icmp_seq 17
64 bytes from 52.85.40.44: icmp_seq=18 ttl=239 time=67.255 ms
64 bytes from 52.85.40.44: icmp_seq=19 ttl=239 time=66.977 ms
64 bytes from 52.85.40.44: icmp_seq=20 ttl=239 time=67.420 ms
64 bytes from 52.85.40.44: icmp_seq=21 ttl=239 time=67.968 ms
^C
--- d2h67oheeuigaw.cloudfront.net ping statistics ---
22 packets transmitted, 15 packets received, 31.8% packet loss
round-trip min/avg/max/stddev = 66.691/67.558/69.138/0.574 ms
~ $

Anybody else seeing these speed issues? It could show up as a whole lot of different sites being weirdly slow, Cloudfront's used by a lot of different companies.