Bitcointalk: Error 504: Gateway time-out [Update: December 2, 2023]

[taufik123]

Yes it happened again a while ago at 2023:11-02 06:05:00 UTC.  So it's happening frequently and we have quite a pattern, usually occurring right after 06:00 UTC

I guess theymos need to check what's going around 06:00 UTC .

And it happened again even at almost the same time as yesterday.
Could this be a network anomaly or even some repair from the central server that somehow always happens at the same time with different durations.

Until now @theymos is still not active and there has been no response from theymos.

Tomorrow we should all be prepared to see if this error will happen again or not.

[1714257037]

1714257037

[1714257037]

1714257037

[1714257037]

1714257037

[NotATether]

And it happened again even at almost the same time as yesterday.

That definitely means it's an automated maintenance work, like a cronjob, or system updates, or the like.

Does anyone know what is the hosting provider for Bitcointalk? Maybe @Cyrus can help out with this since he's in charge of the hosting (I guess?). Or maybe this operation is self-hosted so they can look into this directly.

[WatChe]

Once again Bitcointalk was inaccessible to him and showed Error 504: Gateway time-out.
         

After 10 days we have same issue for almost over an hour.

[Plaguedeath]

Another short time attack.

[taufik123]

And it happened again at almost the same time with more than 1 hour.

After 10 days we have same issue for almost over an hour.

Another short time attack.

This seems to be a fairly routine moment these days.

Cloudflare hosts are always having issues, and downtime happens during @theymos' sleeping hours.

The only thing that needs to be done is to contact the Hosting provider and make sure that the
- The origin server can respond to hostname and domain requests in visitor URLs that generate 502 or 504 errors.
- Investigate excessive load and server failures or network outages.
- Identify applications or services that are timed out or blocked.
https://developers.cloudflare.com/support/troubleshooting/cloudflare-errors/troubleshooting-cloudflare-5xx-errors/

I now also have an Alert for Bitcointalk Downtime and this will monitor what happens when downtime occurs.

[UmerIdrees]

Another short time attack.

I do not think it was a short time as i was trying to access the site for quite some time but to no avail.

Once i knew there was some issue, i kept on refreshing on my mobile. Same Error code 504 as mentioned by the OP.

[NotATether]

And it happened again at almost the same time with more than 1 hour.
This seems to be a fairly routine moment these days.

Cloudflare hosts are always having issues, and downtime happens during @theymos' sleeping hours.

The only thing that needs to be done is to contact the Hosting provider and make sure that the
- The origin server can respond to hostname and domain requests in visitor URLs that generate 502 or 504 errors.
- Investigate excessive load and server failures or network outages.
- Identify applications or services that are timed out or blocked.
https://developers.cloudflare.com/support/troubleshooting/cloudflare-errors/troubleshooting-cloudflare-5xx-errors/

I now also have an Alert for Bitcointalk Downtime and this will monitor what happens when downtime occurs.

FFS. 

It's happening at 6:00 AM UTC every time. So maybe Theymos can get up at midnight once and check all webhosting scripts and/or cronjobs which are talking to the DB to find the culprit.

[LoyceV]

It's happening at 6:00 AM UTC every time. So maybe Theymos can get up at midnight once and check all webhosting scripts and/or cronjobs which are talking to the DB to find the culprit.

Theymos wrote this:

It seems to always happen at around the same time, so I suppose it's probably some automated maintenance process messing something up, but I don't have anything running at this time, so I wonder if it's something in the web-host's networking infrastructure, or something like that. It's difficult to diagnose, since it happens rarely, and usually when I'm sleeping.

I'll use my deleted posts archive to see gaps between posts, starting at post 63 million (October 17). I'll only show gaps larger than 3 minutes between posts (which could mean the server was inaccessible, but it could also mean there simply were no new posts for 3 minutes).

186 seconds between posts at 6300/63006744.html Tue Oct 17 00:08:26 UTC 2023
186 seconds between posts at 6300/63006749.html Tue Oct 17 00:17:02 UTC 2023
246 seconds between posts at 6300/63006776.html Tue Oct 17 00:42:50 UTC 2023
216 seconds between posts at 6300/63006959.html Tue Oct 17 02:07:19 UTC 2023
216 seconds between posts at 6300/63007520.html Tue Oct 17 05:34:25 UTC 2023
234 seconds between posts at 6301/63012218.html Wed Oct 18 00:22:49 UTC 2023
198 seconds between posts at 6301/63017579.html Wed Oct 18 22:06:37 UTC 2023
186 seconds between posts at 6301/63018064.html Thu Oct 19 00:02:49 UTC 2023
259 seconds between posts at 6301/63018293.html Thu Oct 19 01:45:32 UTC 2023
187 seconds between posts at 6301/63018547.html Thu Oct 19 03:50:50 UTC 2023
187 seconds between posts at 6302/63023712.html Fri Oct 20 00:52:02 UTC 2023
229 seconds between posts at 6302/63024251.html Fri Oct 20 04:21:56 UTC 2023
359 seconds between posts at 6302/63028944.html Sat Oct 21 00:42:55 UTC 2023
186 seconds between posts at 6302/63028967.html Sat Oct 21 01:01:20 UTC 2023
270 seconds between posts at 6302/63029016.html Sat Oct 21 01:38:08 UTC 2023
282 seconds between posts at 6303/63033797.html Sun Oct 22 01:44:31 UTC 2023
216 seconds between posts at 6303/63033842.html Sun Oct 22 02:16:13 UTC 2023
215 seconds between posts at 6303/63034137.html Sun Oct 22 04:37:55 UTC 2023
258 seconds between posts at 6303/63038660.html Mon Oct 23 00:17:25 UTC 2023
198 seconds between posts at 6304/63049868.html Wed Oct 25 01:16:19 UTC 2023
222 seconds between posts at 6305/63055527.html Thu Oct 26 00:10:50 UTC 2023
198 seconds between posts at 6305/63055614.html Thu Oct 26 01:09:49 UTC 2023
186 seconds between posts at 6305/63055827.html Thu Oct 26 03:04:44 UTC 2023
186 seconds between posts at 6305/63055901.html Thu Oct 26 03:47:56 UTC 2023
205 seconds between posts at 6306/63060564.html Fri Oct 27 01:12:20 UTC 2023
275 seconds between posts at 6306/63060578.html Fri Oct 27 01:22:25 UTC 2023
204 seconds between posts at 6306/63060666.html Fri Oct 27 02:04:43 UTC 2023
594 seconds between posts at 6306/63060861.html Fri Oct 27 03:54:43 UTC 2023
239 seconds between posts at 6306/63065809.html Sat Oct 28 02:07:43 UTC 2023
341 seconds between posts at 6307/63070434.html Sun Oct 29 00:54:31 UTC 2023
181 seconds between posts at 6307/63070449.html Sun Oct 29 01:08:20 UTC 2023
289 seconds between posts at 6307/63070585.html Sun Oct 29 02:43:38 UTC 2023
269 seconds between posts at 6307/63075227.html Mon Oct 30 00:09:37 UTC 2023
186 seconds between posts at 6307/63075426.html Mon Oct 30 01:55:37 UTC 2023
187 seconds between posts at 6308/63080180.html Mon Oct 30 22:50:02 UTC 2023
209 seconds between posts at 6308/63080470.html Tue Oct 31 00:58:43 UTC 2023
234 seconds between posts at 6308/63085944.html Wed Nov  1 03:05:07 UTC 2023
193 seconds between posts at 6308/63086275.html Wed Nov  1 05:25:32 UTC 2023
2334 seconds between posts at 6308/63086374.html Wed Nov  1 06:44:44 UTC 2023
198 seconds between posts at 6309/63091070.html Thu Nov  2 01:20:08 UTC 2023
311 seconds between posts at 6309/63091674.html Thu Nov  2 06:06:43 UTC 2023
181 seconds between posts at 6309/63091884.html Thu Nov  2 07:05:02 UTC 2023
205 seconds between posts at 6309/63096299.html Fri Nov  3 00:44:38 UTC 2023
228 seconds between posts at 6309/63096436.html Fri Nov  3 02:00:31 UTC 2023
185 seconds between posts at 6309/63096459.html Fri Nov  3 02:15:25 UTC 2023
186 seconds between posts at 6309/63096462.html Fri Nov  3 02:18:49 UTC 2023
197 seconds between posts at 6310/63101573.html Sat Nov  4 01:24:31 UTC 2023
197 seconds between posts at 6310/63101607.html Sat Nov  4 01:53:37 UTC 2023
204 seconds between posts at 6310/63101929.html Sat Nov  4 04:24:55 UTC 2023
276 seconds between posts at 6310/63106327.html Sun Nov  5 00:13:01 UTC 2023
329 seconds between posts at 6310/63106437.html Sun Nov  5 01:11:37 UTC 2023
264 seconds between posts at 6310/63106462.html Sun Nov  5 01:34:19 UTC 2023
193 seconds between posts at 6311/63111491.html Mon Nov  6 00:35:02 UTC 2023
227 seconds between posts at 6311/63111755.html Mon Nov  6 02:59:31 UTC 2023
185 seconds between posts at 6313/63132878.html Fri Nov 10 00:43:49 UTC 2023
185 seconds between posts at 6313/63132951.html Fri Nov 10 01:14:43 UTC 2023
6041 seconds between posts at 6313/63133565.html Fri Nov 10 07:45:31 UTC 2023

Most of those are at UTC night, but Ninjastic Space's graph confirms users are less active at that time, so there could be many false positives in my list.

Note to self for updates: see find_gaps_between_posts_to_track_server_downtime.txt

[SmartGold01]

I encountered it again around 8 to 9 am my local time even as I was active around 1 am to 4am forum time I didn't experienced any downtime and this could probably be some host error, although I don't usually panicked anymore because I know too well after 20 to 40 minutes times it will get cleared off. And I want to asked is there no alternative way to get this error finally put to an end? Sometimes it's very annoying when posting and one encountered such error sometime I do lost my composed post, and again whenever you writing a post without previewing yet and you mistakenly lost the post it's very hard to create back same quality of the post you already lost.

[Bureau]

I might be wrong or it might be my ignorance. Isn't this issue of seeing the forum down became more prevalent after the Bitcoin mixer issue. I mean it is more frequent and long lasting before the mixer got down by the FBI. I think there is something wrong and I feel the admin should have two urls. One .org and the next should be .onion.

[decodx]

I might be wrong or it might be my ignorance. Isn't this issue of seeing the forum down became more prevalent after the Bitcoin mixer issue. I mean it is more frequent and long lasting before the mixer got down by the FBI.

Which mixer are you referring to? You need to be more specific, given that there have been multiple mixer sites shut down by the FBI/Europol.
But, I fail to see the relevance to the issues with the bitcointalk hosting server.

I think there is something wrong and I feel the admin should have two urls. One .org and the next should be .onion.

And what if the cause of the problem lies on the hosting side, as theymos has already mentioned? How will having a .onion mirror domain help in that case?

[PX-Z]

I might be wrong or it might be my ignorance. Isn't this issue of seeing the forum down became more prevalent after the Bitcoin mixer issue. I mean it is more frequent and long lasting before the mixer got down by the FBI. I think there is something wrong and I feel the admin should have two urls. One .org and the next should be .onion.

This mere forum has nothing to do about those shutdown mixers. Yes, they can advertise here, but nothing more than that.

Since this forum's hosting server(namecheap) is based on US and if they wants to summon theymos to the court for some legal reasons they do that already if there is, and they can just request to namecheap to shutdown the site's server for good. But that:s not the case. As per theymos, the downtime is almost the same hours from previous ones, and conclude its just a scheduled maintenance from the hosting, nothing more.

[logfiles]

I now also have an Alert for Bitcointalk Downtime and this will monitor what happens when downtime occurs.

Which website/service does push out alerts like this if you don't mind sharing.

I might be wrong or it might be my ignorance. Isn't this issue of seeing the forum down became more prevalent after the Bitcoin mixer issue. I mean it is more frequent and long lasting before the mixer got down by the FBI. I think there is something wrong and I feel the admin should have two urls. One .org and the next should be .onion.

Lol. This has nothing to do with Bitcoin mixers. DDoS attacks or downtimes can have to any site, including big ones like Twitter or Facebook. If the Law enforcement authorities were unhappy with how mixers are being advertised in this forum. They would have already contacted admin by now or even completely seize the forum and shut it down, which wouldn't be hard for them. You are just over imagining.

[taufik123]

-snip-
As per theymos, the downtime is almost the same hours from previous ones, and conclude its just a scheduled maintenance from the hosting, nothing more.

Will it always be scheduled maintenance without any prior confirmation?
There should be an explanation for all these downtime events, as some people are starting to speculate about strange things.

Which website/service does push out alerts like this if you don't mind sharing.
-snip-

This is just a kind of monitoring website that can do real-time testing.
The name is Pingdom
a tool used to test websites.

Pingdom Tools itself operates online and is also freely available to
anyone who wants to try it (30-day trial).
Pingdom provides valid analysis reports
related to load time, page size, website structure details, and total number of HTTP
requests.

You will also receive email alerts about website uptime, website speed, and more.

[joker_josue]

This is just a kind of monitoring website that can do real-time testing.
The name is Pingdom
a tool used to test websites.

Pingdom Tools itself operates online and is also freely available to
anyone who wants to try it (30-day trial).
Pingdom provides valid analysis reports
related to load time, page size, website structure details, and total number of HTTP
requests.

You will also receive email alerts about website uptime, website speed, and more.

The uptime that appears in the image is related to what time interval? A month, year, week?
If in the last year the uptime is below 90%, it is really not a good sign.

We hope that the planned renewal of the servers will resolve most of the problems that occurred.

[taufik123]

The uptime that appears in the image is related to what time interval? A month, year, week?
If in the last year the uptime is below 90%, it is really not a good sign.

The reading on Pingdom only starts when I add bitcointalk.org to the Pingdom uptime list.
So the uptime below 90% is the interval from the first time I added it which was when Donwtime was last yesterday.

But it will keep running and will tell how the website is working and will get email alerts when downtime occurs.
But yes this is only valid for 30 days trial and have to renew for next month to keep getting premium features.

We hope that the planned renewal of the servers will resolve most of the problems that occurred.

The server update is completely done by the administrator, we just need to wait for @theymos to fix everything.
But there is no recent statement from @theymos about the downtime problem that always occurs and how to follow up.

[PX-Z]

-snip-
As per theymos, the downtime is almost the same hours from previous ones, and conclude its just a scheduled maintenance from the hosting, nothing more.

Will it always be scheduled maintenance without any prior confirmation?
There should be an explanation for all these downtime events, as some people are starting to speculate about strange things.

I can't answer to that, and you're right, well, even theymos just think it that way, we dont know what he already did to fix this or if there is need to fix since we didnt heard him yet.

[salad daging]

Just now the forum error again showed as the OP said but it did not take long only a few seconds or approximately 1 minute, but I see lately the forum often has problems, what is really happening?

Just now you felt the same thing?

[taufik123]

Just now the forum error again showed as the OP said but it did not take long only a few seconds or approximately 1 minute, but I see lately the forum often has problems, what is really happening?

Just now you felt the same thing?

Yes, the forum crashed again for exactly 3 minutes and then came back online.

Looking at the Result Log the Downtime locations are Amsterdam, Madrid, Oslo and Seattle 2.



The Amsterdam server I see often experiences downtime compared to other servers.

Outage analysis of bitcointalk.org

[philipma1957]

I have been on this forum since 2012. I have seen 100 outages in the 11 years I have been on this site.

There always seems to be more down time when BTC starts to bull run.