1

We are running an app and website on Google Cloud servers and beginning a couple of days ago users running on a specific internet service provider (Time Warner Cable) are experiencing very slow loading times, time out messages, and other problems communicating with our IP / URL.

We contacted Time Warner Cable and they are not providing much assistance.

We do not have a support plan currently with Google so we are not able to get specific information on this issue from them.

We did not make any changes on our side around this. How can we identify why there is slowness through this specific ISP?

Are there ways to gather information around this other than running a traceroute? Is there information within the traceroute results that can help us identify the cause of this issue?

6
  • Is there information within the traceroute results that can help us identify the cause of this issue? - Yes there is. Do you see latency at any particular node that carries throughout the path? Also, you can use MTR, WinMTR, pathping, SmokePing, etc. to analyse the path.
    – joeqwerty
    Nov 15 '16 at 19:03
  • @joeqwerty thanks for the additional info, it appears there is some delay at certain points in the traceroute, but that is also the case for a traceroute run on a device connected via a different network that is not experiencing the same slowness
    – Dan
    Nov 15 '16 at 21:14
  • Right, the key is to determine if that delay carries through the path. If you see delay at a particular hop but not at any subsequent hops then that hop is probably not affecting the path. If you see delay at a particular hop and that delay carries through subsequent hops then that hop is probably affecting the path. It's all about knowing whether a particular hop is merely ignoring/low prioritizing your ICMP traffic or if it's actually dropping legitimate traffic. Seeing delay that carries through the path is usually a good indication that there's a real problem in the path.
    – joeqwerty
    Nov 15 '16 at 21:25
  • If the issue is just affecting users reaching through a particular ISP, the resources in your project must be running fine. I think the approach suggested by @joeqwerty would be the right one. I would only add to perform the tests in both ends (Problem location to Google and vice versa). Packets returning from Google might follow a different path than the incoming ones.
    – Carlos
    Nov 16 '16 at 19:59
  • Were you able to solve this issue? If this is the case, you might consider posting an answer, so the community can benefit from your experience.
    – Marilu
    May 24 '17 at 20:29

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.