Since 26 Jan, 2015 particular websites have suddenly become unusable during peak hours (FiOS 75/75, Boston are). Speedtests show all is well with the connection and accessing the site(s) via VPN is nice and fast. Traceroutes show a variety of different different routes with varying degrees of goodness. Pings don't show any packet loss. Presumably this is a peering issue, though at this point I'm well beyond my understanding. So, the question is How do I debug this issue and what information can I collect to help the techs actually fix the problem.
FWIW, one of the sites in question is dailykos.com. They seem to be using fastly CDN for much of their content.
Any guidance would be most welcome.
Thanks
PS: speedtest just now while the example site will not load on FiOS and fully loads in a second over VPN:
[att=1]
Oh, and here's a good traceroute illustration:
$ traceroute images.dailykos.com
traceroute to fallback.global-ssl.fastly.net (23.235.46.185), 64 hops max
1 10.1.1.1 0.464ms 0.364ms 0.298ms
2 173.48.161.1 5.223ms 4.793ms 4.904ms
3 130.81.191.250 7.387ms 7.466ms 7.366ms
4 130.81.151.76 4.909ms 4.905ms 4.932ms
5 152.63.20.117 12.441ms 12.433ms 12.434ms
6 152.63.4.141 12.373ms 12.468ms 12.370ms
7 152.179.220.126 14.989ms 14.937ms 14.872ms
8 68.86.84.249 12.505ms 12.259ms 14.988ms
9 68.86.85.25 22.409ms 22.433ms 22.519ms
10 68.86.85.1 19.817ms 19.894ms 19.919ms
11 68.86.83.74 17.389ms 19.870ms 22.460ms
12 4.53.116.18 24.815ms 25.080ms 24.851ms
13 * 208.122.44.198 2849.473ms 29.832ms
14 173.231.134.18 22.430ms * 2026.736ms
↧