A user of my Discourse in Belgium complains our site responds too slowly to clicks. I’m in California. I’ve not gotten any complaint from my other users, who so far are in Minnesota, Germany, and the Netherlands. The user in question tells me she uses Chrome on Windows. Her download and upload speeds test similarly to mine via speedof.me.
If I visit https://discourse.MY_DOMAIN.com/logs/, I see the latest entry is time-stamped just a few minutes before her email was sent. Its “info” tab reads:
User is very tech un-savvy, so I doubt she’s on VPN, but I could ask. I asked her to log in to safe mode and report if that improves anything, but she’s yet to reply to that (it is late in Europe now).
Can you recommend/link a traceroute for Discourse?
Traceroute is a command line tool, present on mac (traceroute) and Windows (tracert). It will show you the network route from their computer to your server, hop by hop. I’m not aware of any good browser-based alternatives, most will show the route from the server rather than the client.
You could ask for a browser screenshot to see if there are any funky extensions installed too.
Have you ruled out local security software? I’ve occasionally seen antivirus browser extensions which proxy traffic and can cause issues.
Ah. So, I can SSH into our Discourse machine, I can see that traceroute isn’t installed, I can run apt install traceroute (installs version 2.1.0), and then I can run traceroute HER_APPARENT_IP_ADDRESS_FROM_DISCOURSE LOGS?
If I do, it outputs the following.
traceroute to 84.196.9.6 (84.196.9.6), 30 hops max, 60 byte packets
1 * * *
2 10.70.134.15 (10.70.134.15) 1.101 ms 10.70.134.35 (10.70.134.35) 1.079 ms 10.70.134.27 (10.70.134.27) 0.988 ms
3 138.197.251.92 (138.197.251.92) 1.323 ms 138.197.251.94 (138.197.251.94) 1.628 ms 138.197.251.124 (138.197.251.124) 1.206 ms
4 138.197.251.110 (138.197.251.110) 1.079 ms 1.071 ms 138.197.251.114 (138.197.251.114) 1.056 ms
5 138.197.244.17 (138.197.244.17) 1.247 ms 1.251 ms 138.197.244.19 (138.197.244.19) 1.112 ms
6 nyk-b3-link.telia.net (62.115.45.5) 1.866 ms 1.383 ms nyk-b3-link.telia.net (62.115.45.9) 1.331 ms
7 * * *
8 ldn-bb3-link.telia.net (62.115.113.21) 82.665 ms 82.486 ms 82.459 ms
9 adm-bb4-link.telia.net (62.115.134.26) 78.418 ms adm-bb3-link.telia.net (62.115.113.210) 82.951 ms 83.025 ms
10 brx-b3-link.telia.net (62.115.116.191) 82.894 ms brx-b4-link.telia.net (62.115.116.231) 78.732 ms 78.321 ms
11 be-dgb01a-rb1-ae-20-0.aorta.net (213.46.162.13) 82.955 ms 82.970 ms be-zav01a-rb1-ae-21-0.aorta.net (213.46.162.6) 87.143 ms
12 * * *
13 * * *
14 * * *
15 * d54C40906.access.telenet.be (84.196.9.6) 94.362 ms 93.255 ms
The longest hop is the last one, 94 milliseconds, which seems reasonable. Would this indicate the route from our Discourse machine in Northern California to her computer in Belgium looks like it should be just fine? Was the Discourse log in my original post above otherwise devoid of any valuable information other than getting her apparent IP address?
My user overnight reported that logging in via Discourse safe mode solved her problem. I’ve now asked her if she can help narrow which of the three Safe Mode checkboxes actually helps.
My challenges continue to be our time-zone difference, our language difference, and our difference in comfort with technology. I can update this thread if/when I learn more.
Thanks also for the relevant news article. This was her first time using our Discourse, so now I also wonder if that outage could be relevant.