Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Meta outage

metastatus.com

Post reply on HN
Print view View original

621–630 of 902 posts

Re: Meta outage

#621

Post by consumer451 »

jon-wood wrote: post #374
DeusExMachina wrote: post #281

Looking at the Downdetector home page [1], it looks like many more services are having outages, not just the ones owned by Meta, including: - Google - YouTube - Google Play - T-Mobile - X (Twitter) - Discord - TikTok - Pokemon Go - Snapchat It looks like they all have the same failure point. [1] https://downdetector.com

Or people are using Facebook Auth for them. I don't really trust Down Detector, which despite the claims is really People Winging on Twitter Detector.

> Or people are using Facebook Auth for them.

Gmail is also on the list. You can't use FB auth to login to Gmail, can you?

Re: Meta outage

#622

Post by AnimalMuppet »

ndiddy wrote: post #535
neocritter wrote: post #499

Earlier quoted context omitted.

It's real. Single core performance improves all the time. People overestimate how much power it takes to handle lots of queries per second on a well-tuned system and well-written software in 2024. https://hn.algolia.com/?dateRange=all&page=0&prefix=true&que...

I see the "sorry, we are receiving too many requests, try again in a few minutes" error several times a day on here. I don't think that HN is reliably able to handle the amount of users it currently has.

Hmm. I get a different message. Something like "We are having trouble handling your request. Sorry!"

I saw it just a few minutes ago, but I don't remember the exact wording...

Re: Meta outage

#623

Post by Turing_Machine »

packetslave wrote: post #567
Solvency wrote: post #478

[flagged]

"blah blah blah I don't like the company so all their engineers must be really stupid..."

Yep. I've also noted that the people making such claims never seem to cite their own work as an example of how to implement something at Facebook or YouTube scale that is less "brittle".

Armchair quarterbacking isn't just a U.S. football phenomenon.

Re: Meta outage

#624

Post by pixl97 »

partiallypro wrote: post #579
DeusExMachina wrote: post #281

Looking at the Downdetector home page [1], it looks like many more services are having outages, not just the ones owned by Meta, including: - Google - YouTube - Google Play - T-Mobile - X (Twitter) - Discord - TikTok - Pokemon Go - Snapchat It looks like they all have the same failure point. [1] https://downdetector.com

Down Detector is so unreliable. People that can't call an AT&T phone via Verizon will think (and report) that Verizon is down, when it's really AT&T. People can try logging in using Facebook's on click login and not be able to get in, so they think Tiktok is down. It's not all that useful. I hate when journalists cite it.

Eh, mostly it's people misunderstanding what it represents.

If I can't login to tiktok because FB is down, then tiktok is effectively down for me. When it comes to technology most people don't care about the trip, they care about the destination.

So yea, tiktok isn't "down" but for a lot of people it might as well be, hence coupling your infrastructure/auth on other providers has side effects like this you must take into account.

Re: Meta outage

#625

Post by ceejayoz »

colpabar wrote: post #537
redserk wrote: post #456

Earlier quoted context omitted.

Yeah, it's wild that it's now treated as an authoritative source, especially by some news organizations. It's as good as asking a neighbor what happened with a loud noise down the street. Sometimes you'll get something good, sometimes it'll be completely wrong.

Downdetector is nice because it answers my question of "is anyone else having issues with this?" When it takes AWS an hour to even acknowledge "increased error rates", and tells me that everything is a-ok in the meantime, I want another perspective.

Twitter's search used to be my go-to for this - a search for "AWS down" would typically be very illuminating - but it's tough to get it to genuinely spit out the most recent tweets with a keyword these days.

Re: Meta outage

#626

Post by laughablecargo »

ak_111 wrote: post #205
laughablecargo wrote: post #193

Earlier quoted context omitted.

In an election between the two parties I think that would be a very likely narrative but in this case it's the primaries. And (as others have mentioned) neither primary has really had anything resembling a competition at this point so it's more or less irrelevant to the result (most likely).

conspiracy theory get more imaginative. For example, you can then pose the conspiracy that the fact it was Russian interference conspiracy was a conspiracy to justify more social media policing during the real election.

Perhaps but why would downtime increase policing? Something like a large scale fake news campaign (i.e. basic deep fakes or something) is easier (generally speaking) to pull off and would be more likely to cause increased policing that a short outage.

Re: Meta outage

#627

Post by ImPostingOnHN »

ceejayoz wrote: post #618
rimunroe wrote: post #551

Earlier quoted context omitted.

> I don't really trust Down Detector, which despite the claims is really People Winging on Twitter Detector. I'm confused. Isn't listening for spikes in complaints about outages a great way to detect them? I know for a fact some service companies monitor social media channels for this purpose (among others). I'd be surprised if that wasn't more or less standard practice. I've checked Down Detector for ISP outages in…

> Isn't listening for spikes in complaints about outages a great way to detect them? When there's a major ISP outage, people report problems with all the major sites. When Facebook's down, people report problems with any site that has "Login with Facebook" as an option. It's almost never actually an outage impacting all of FAANG at once.

> When Facebook's down, people report problems with any site that has "Login with Facebook" as an option.

If users log into your site with Facebook, then the login functionality of your site effectively is down when "Login with Facebook" is down.

From the user's perspective, your subcontractors, including authentication subcontractors, are a problem for you to deal with and never show them. From your perspective, you could have architected your site in a way that logging in doesn't "go down" when Facebook login is down.

If the user chooses "Login with Facebook" over other authentication options available, and they don't want to use other options, educating them with a good error message might help. Or you could remove the Facebook login option, if you (totally reasonably) don't want Facebook's failures to reflect poorly on you.

Re: Meta outage

#628

Post by baby_souffle »

ceejayoz wrote: post #618
rimunroe wrote: post #551

Earlier quoted context omitted.

> I don't really trust Down Detector, which despite the claims is really People Winging on Twitter Detector. I'm confused. Isn't listening for spikes in complaints about outages a great way to detect them? I know for a fact some service companies monitor social media channels for this purpose (among others). I'd be surprised if that wasn't more or less standard practice. I've checked Down Detector for ISP outages in…

> Isn't listening for spikes in complaints about outages a great way to detect them? When there's a major ISP outage, people report problems with all the major sites. When Facebook's down, people report problems with any site that has "Login with Facebook" as an option. It's almost never actually an outage impacting all of FAANG at once.

> It's almost never actually an outage impacting all of FAANG at once.

Exactly. If you click through down detector when things are _up_ you'll see people still complaining that $site is down. Could be a local power outage or even a flaky connection in their own home.

Down Detector is one of many signal sources and should have a "credibly" score associated with it that's proportional to the number of people complaining that something's down.

Re: Meta outage

#629

Post by jordanthoms »

ceejayoz wrote: post #618
rimunroe wrote: post #551

Earlier quoted context omitted.

> I don't really trust Down Detector, which despite the claims is really People Winging on Twitter Detector. I'm confused. Isn't listening for spikes in complaints about outages a great way to detect them? I know for a fact some service companies monitor social media channels for this purpose (among others). I'd be surprised if that wasn't more or less standard practice. I've checked Down Detector for ISP outages in…

> Isn't listening for spikes in complaints about outages a great way to detect them? When there's a major ISP outage, people report problems with all the major sites. When Facebook's down, people report problems with any site that has "Login with Facebook" as an option. It's almost never actually an outage impacting all of FAANG at once.

That's kinda the point though isn't it? DownDetector is showing an early indication of a major outage in both of your examples. The issue may not be caused by the indicated service, but it's still a useful information source especially when we can correlate reports on there with what we are seeing in our internal monitoring.
Post reply on HN