12

SO status

SO doesn't seem to allow logging in or writing anything, but status page (https://www.stackstatus.net/) says everything is operational. Did I miss something?

8
  • OP: if you don't mind, could you add the approximate time when this happened, ideally in UTC but with a non-UTC timezone works too?
    – cocomac
    Oct 30, 2023 at 4:35
  • I noticed this happening on a few sites about three hours ago, at about: 2023-10-30 1:30:00Z Oct 30, 2023 at 4:39
  • @cocomac - I can confirm, I'm getting the same on Arqade (Gaming). Intermittent Read-only AND full "offline" messages from about 23:00 UTC until now (04:40 UTC)
    – Robotnik
    Oct 30, 2023 at 4:39
  • raspberrypi.stackexchange.com & raspberrypi.meta.stackexchange.com Mon 2023-10-30 06:12:23 UTC "This site is currently in read-only mode. We’ll return with full functionality soon. Visit our status page or search our recent meta posts on the topic for more info." has been like this for > 4 hours
    – Milliways
    Oct 30, 2023 at 6:14
  • 2
    I was seeing the same problem (although it's gone away at the moment) and it seemed to be related to using Starlink. As soon as I routed stackoverflow through my other gateway (Optus) I had no issues at all.
    – Nick
    Oct 30, 2023 at 7:43
  • 2
    No issues here. Sounds like it may be related to a specific DNS and/or ISP? I had similar problems long time ago and then it was caused by DNS.
    – Lundin
    Oct 30, 2023 at 10:11
  • The site was in read-only mode for me here in Sydney, so at least from 2PM to 6PM Sydney time on October 30, which would be 3AM to 7AM UTC on October 30, which would have been from 11PM to 3AM New York time. Oct 30, 2023 at 22:48

1 Answer 1

13

The application for Stack Overflow encountered an error that caused the site to temporarily become unavailable from our primary data center (New York), as a result traffic was automatically diverted to the disaster recovery site (Colorado) which runs in read only mode. Before moving to Cloudflare, an error like this would have caused a maintenance page to load instead of the read only site.

We're currently investigating the root cause of the original error.

10
  • Also, looks like it did not affect all users, how can it be? Oct 30, 2023 at 20:10
  • 5
    @ShadowWizardIsSadAndAngry I suspect the site was actually flapping which meant some users were able to load while others were directed to the DR site.
    – Josh Zhang StaffMod
    Oct 30, 2023 at 20:27
  • Thanks @Josh. Would be somewhat ironic in the root cause of the original error is stack overflow error in some code. ;-) Oct 30, 2023 at 21:20
  • 4
    So there are two issues here. Both the original error which caused the site to become unavailable, and the fact that Stack Status continued to report that everything was operational the entire time. Don't forget that the OP was more concerned with the second of these two issues, but this answer only addresses the first issue. Oct 30, 2023 at 23:02
  • 1
    @BrettDonald in my experience, StackStatus never reflects uptime problems- at least- never while they are actually happening. I have stopped being surprised. In fact, I would be surprised if it actually did its job for once.
    – starball
    Oct 31, 2023 at 0:07
  • 1
    StackStatus is tied into our incident management software and isn't tied to our uptime monitoring software. It's manually updated when an incident is declared.
    – Josh Zhang StaffMod
    Oct 31, 2023 at 7:11
  • 1
    @JoshZhang You might want to answer this question about it that came up today with that info. Oct 31, 2023 at 8:21
  • 2
    @JoshZhang Is there any reason (technical or otherwise) why StackStatus wouldn't or couldn't be tied into the uptime monitoring software? Oct 31, 2023 at 15:24
  • 2
    @CaveJohnson Well, I think there is some granularity and/or precision missing, e.g. "up but read only" isn't the same type of outage as "down for everyone" or "unreachable to some portion of the world." There are probably things the status monitor could do to determine the difference, but if previous conversations are any indication, changes are hard. In the meantime, since updates are manual, not all "outages" (e.g. geographical ones) will appear. It has been mentioned before: if you have to pick, do you want the on-call person at 2AM fighting the fire, or updating the status page? Oct 31, 2023 at 17:15
  • 1
    @Stuckat1337 I'm fine if StackStatus stays what it is. But I want to the server error page to make it clear what I'm going to see when I go there- that it is likely to lag behind in reporting what's going on right now, how long on average it's going to lag behind, and why it lags behind.
    – starball
    Oct 31, 2023 at 19:55

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .