I’m only having issues with the SDMB. Other sites load just fine, same as always. Maybe a message board such as this, which involves interaction and uploading as well as downloading, reveals connection issues that sites I just download do not, but…
I’ve tested my connection with speedtest[dot]net twice this morning. Both times the download speed was well over 400Mbps and upload was over 40Mbps. I suppose my connection could have some sporadic wonkiness issues that don’t necessarily show up in an internet speed test, but…
Other Dopers are reporting similar issues in this thread.
I haven’t seen them today (yet) but there were absolutely problems yesterday that were isolated to this forum. No problem with any other site or any other program or device that connects to the internet.
Just corroborating: I’ve been noticing much the same issues as you and @kenobi_65 have reported in this thread.
It’s not been all that frequent – maybe once in a day, and then it usually won’t happen for another day or two. Call it three or four times a week. Been this way for roughly two months.
EDIT: And as @Atamasama just posted: When I notice an issue with the SDMB loading, other websites I have open in other tabs are functioning normally.
And FWIW, I only notice these issues while surfing via Safari on my iPhone (latest version of iOS). On Win 11, Chrome browser (automatically most recent version), I’ve never seen a real issue.
It’s probably premature to draw conclusions like that since the problems are indeed sporadic and fairly infrequent. But they’re there. Just today (with the Edge browser) I got the rotating circle when trying to scroll up in a thread for such a long time (several minutes!) that I gave up, shut down the browser, and restarted. The problem then went away. There may be some browser-related issue but this never happened before and just started approximately a week ago.
I realize this kind of vague symptom may not be helpful to the developers, but I’m just reporting what I’m seeing. If there’s any message I want to convey to the development team, it’s that when a site is working well, sometimes the best policy is to just leave it the hell alone. I know that’s very hard to do when your whole job centers around continuing to change things, but dammit, sometimes it really is the best policy.
ETA: That last paragraph was added as an edit. I thought the edit had been lost because it took more than a full minute to show up!
It’s possible that this whole thing may be due to a reallocation of loads to different servers and not a software problem at all, but, again, there’s no doubt that there’s definitely a recent issue that isn’t on the client side.
It’s hard to precisely troubleshoot these issues because the SDMB isn’t its own server/machine. It runs on shared Discourse infrastructure, which is part of shared Amazon infrastructure. As far as I know, our mods can’t really see the resource usage or hardware limits of this particular board; it’s all hosted and transparently handled by the company that makes Discourse. It’s all just in a virtual machine running in the cloud running amongst thousands (or probably hundreds of thousands) of other VMs in a huge shared regional Amazon data center.
It’s entirely possible for there to be regional network issues/congestion, or problems with the particular shared Amazon machines that the SDMB runs on. In those cases, switching computers or browsers won’t make a difference. Speed tests and other websites won’t really isolate the problem either, beyond telling you that it’s not an issue with your particular device. It may still be a problem between your ISP and the SDMB’s particular instance of Discourse.
But you can try a different network (like your phone hotspot or the Cloudflare VPN) or open your browser console and see what might be taking abnormally long… nothing should take more than a couple seconds under normal conditions, and usually way less than that. If you find something that’s taking 5000ms+, there’s your culprit.
Have the admins confirmed this? I was under the impression the Dope installed the software on their own servers similar to how the old software was managed.
It’s probably also why we are tied to their hosted software release cadence (which some of our members struggle with). At work, we run our own self-hosted Discourse instance and have full control over everything. By contrast, I think the SDMB mods/admins can only manage a few plugins and admin settings through the Discourse admin interface, but not the underlying software or network stacks. (Of course, if I’m wrong, hopefully one of the mods will be along to correct me shortly.)
Anyway, if this is correct and we are indeed hosted and managed by Discourse*, it just means we have limited control over technical issues experienced by the forum. That’s up to the company and AWS. But you’d hope AWS would do a better job of it than our non-profit owners that don’t even remember we exist… lol. It really is such an interesting arrangement we have with them.
(*Or rather, the company behind them… I always forget its name, something like Civilized Discourse Construction Kit or something… sounds like something out of the Fallout games…)
This makes this annoying and persistent YouTube links bug all the more maddening. I had sort of given it a pass since I assumed a fix would require our volunteer IT department to manually apply a major version update and shut everyone out for maintenance and the inevitable upgrade bugs. But this makes it pretty clear that it’s Discourse that has continued to screw that specific goat.
The fact that we’re hosted by Discourse is the reason we have no control over the 10,000 post limit per thread, one of probably many performance-related parameters we don’t control. Those who self-host can set the limit to be whatever they like.
