Thread 106949883

>AWS goes down
>A good chunk of the internet goes down with it

Uh, nice centralized eco-system you got there, ameriburgers. Pretty sure this is one of the first things you're taught NOT to do when setting up a network

AWS has tons of competitors in the cloud space now, not a good look for them

>>106949883 (OP)
Yes, because not a single company in another country has ever experienced an internet outage.

>>106949891
The cost of switching is too high and no one was ever fired for choosing AWS

>>106949883 (OP)
I bet 10 of your monopoly money currency that your shithole's companies use AWS

>>106949941
You owe me 10 bucks (im unemployed).

Its not even the entirety of aws, its just us-east-1. One region goes down and half the internet goes down because everyone wants us-east-1 since the newest features are rolled out to us-east-1 first and everyone subsequently wants the newest features on that region so the others are underutilized relative to use-east-1 usage.

t. former AWS engineer

>all it would take to bring down the internet is damaging Amazon's servers
why haven't terrorists taken advantage of this vulnerability yet?

>>106949883 (OP)
And nothing of value was lost.

>>106949883 (OP)
but how will zhey problem-reaction-solution ze cyberpandemic if we don't do things this way?

>>106950368
>why haven't terrorists taken advantage of this vulnerability yet?
I mean noting really important is hosted in AWS as far as I know, the pic of OP only shows inane crap and the worst thing I've heard so far was a few ATMs not working

AWS didn't go down. It was a DNS error at a data center.

>>106951378
Same thing really.

>>106949883 (OP)
Use a VPS instead. I have rented a VPS on Hetzner for the last 7 years and it has worked fine.

>>106950368
terrorists goon online too

>>106950368
how are they going to do that?

>>106951378
>DNS error
DNS still shitting up systems in 2025. based.

>>106949891
This. They were dominating in 2015-2020, since Covid microsoft azure is the biggest competitor.

>>106950368
because terrorists don't actually exist and terrorism never happens unless being funded but the government

>>106950322
Same here, I'm also a 29 year old virgin who spends his entire life on the web.

>>106949891
>tons
I only know of microsoft and oracle

>>106949883 (OP)
>>AWS goes down
>4chan still working
Based

>The incident highlights the complexity and fragility of the internet, as well as how much every aspect of our work depends on the internet to work,” Daoudi said in a statement to CNN.

I like how they keep dancing around the fact it is the fucking cloud, not the goddamn internet that was the problem.

>>106952820
>Internets
I hate the cloud. I would hope that this would mean people would move away from the cloud, but it will never happen. Instead dipshit CEOs will just be "internets" thanks to CNN and the other shitheads that run the media.

>>106949883 (OP)
>it's been over 12 hours and if anything it's worse than 8 hours ago
I'm gonna take a fat shit on my CTO's desk, who sperged out so much about m-muh hetzner shitbox reliability we must move to AWS and we spent like two months on this shit for absolutely no reason and 5x higher spend

>"Internet is on the fritz"
>it's actually just a bunch of companies who sold their soul to Amazon
lmao

Even 4chins is slow.

>>106950368
it's honestly impossible in AWS's case. just to take down a single region, you'd have to target every availability zone in the area. each AZ has critical components of it isolated, so if you destroy one AZ the idea is that all the others can continue to function independently.
so maybe you have reliable information on the architecture of an AWS region and have access to 3+ commercial jets that you will kamikaze with. that's just a single region and nearly every company that is on the cloud is replicating their infrastructure in at least 2 regions (of which, the other may not even be in the same country)

>>106953170
Kind sir please understand it is Diwali

>>106953213
>it's honestly impossible in AWS's case
are you fucking serious RIGHT NOW half of AWS is down globally (or rather most things are "up" but you basically roll the dice whether anything works or not) because us-east-1 is fucky
us-east-1 has been aws' single point of failure for over a decade
>that's just a single region and nearly every company that is on the cloud is replicating their infrastructure in at least 2 regions
I have a bunch of shit on aws tokyo and it's been fucked for about 12 hours now

>>106953292
what service disruptions did you face in aws tokyo apart from global services?

>13 hours later
>97/139 services still fucked
I have a pair of $200 mini PCs, where if one fails the second takes over. If internets go down it switches to mobile, and if power goes out there's a fatty 1.5kW power station which lasts like 2 days for those tiny devices
I think I have more 9's in reliability than AWS lmao

>>106949883 (OP)
>AWS goes down
>A good chunk of the internet goes down with it
Somehow nothing of value was affected

>>106954177
think of the shareholders anon.

>>106954177
>Impacted (97 services)
kek it was ~40 when I checked several hours ago
we must accelerate

>>106954177
wtf are all these things man....how did it come to this

>>106954679
What are all the monkeys doing to fix this?

>>106954803
>What are all the monkeys doing to fix this?
Bloody benchod, is celebration now

>>106950368
Nuking northern Virginia would take down the ENTIRE internet.

>>106953292
After looking into it you are right. I thought aws of all places had contingencies in place for this sort of thing but I guess they just dropped the ball on this heavily

>>106950368
>anon finally watched Mr Robot

>>106949883 (OP)
>Pretty sure this is one of the first things you're taught NOT to do when setting up a network
It seems you have zero clue on how cloud based software works, then.

>>106949883 (OP)
>They could hold the world for ransom if they wanted
>nu/g/ doesn't care as long as they have AI toys
Pic unrelated. I just wish I could cuddle with seal Miku.

>>106952223
so close, still based

>>106952389
sounds easy, you can just blame everything on AWS

>>106949883 (OP)
The outage is over. You lose again, chuds.

Wait sorry dumb question but did this only affect North America?

I was browsing reddit and using Canvas during my MBA at the time. Canvas lagged a bit but was perfectly fine.

>>106949883 (OP)
>ameriburgs
um excuse me the pc term is "burgoid"

>>106950368
Because they literally don't exist. Some of the biggest hacks have been kids scamming a few hundred bucks lately, that's not even a joke.

>>106950368
it is not a vulnerability, it's a feature. to be used in key moments

>>106949891
>>106952193
>>106952588
lmao
if jeetsoft was THE competition, AWS can take a hit or two just fine.
jeetsoft lives on bureaucracy, not meritocracy. and that was always true, even before it got fully enjeetified.
the real meritocratic competition comes nowadays from chinks along side the usual competent options like cloudflare.

>>106952620
Powered by Rust(™)

>>106952620
Not even Reddit was safe. 4chan is truly the superior website.

>>106950374
roblox down is positive value actually

>>106958148
>jeetsoft lives on bureaucracy, not meritocracy. and that was always true, even before it got fully enjeetified.
do we know if they are they kept in the windows playpen while the experienced microsoft devs all work on cloud infrastructure now?
azure seems HALFWAY competent compared to windows

>>106954177
>If internets go down it switches to mobile, and if power goes out there's a fatty 1.5kW power station which lasts like 2 days for those tiny devices
Giga based failover master. If I ever self host I aspire to be like you.

i did not even notice it. Wasn't for a thread in /g/, I would not have found out until much later.

Now imagine if chinks and ruzzkies attacked US servers of Oracle, Google, Apple at the same time.

Telling you, we need to create more independent data servers without any chinese hardware and software in it. The war will start together with the cyber attacks 100%.

>>106959431
>do we know if they are
yes, and objectively so. take github for example, which has pre-jeetsoft and jeetsoft lifes.

>In the two years since the acquisition announcement, GitHub has reported a 41% increase in status page incidents. Furthermore, there has been a 97% increase in incident minutes, compared to the two years prior to the announcement.
>Before the acquisition, there were 89 incidents published on the GitHub status page. After, there were 126 incidents. A 41% increase
>In the 24 months prior to the acquisition announcement, there were 6,110 minutes of downtime. During the 24 months after, there were 12,074 minutes of downtime, a 97% increase

https://web.archive.org/web/20200929202120/https://statusgator.com/blog/2020/06/04/has-github-been-down-more-since-its-acquisition-by-microsoft/

of course the authors above had to do some spin for jeetsoft because it's not good to anger big corpos who are also potential customers/investors/buyers. why that blog post doesn't exist anymore, i don't know.

>>106949883 (OP)
I've never used anything from that list. Why do people use slop?

>>106961364 (Me)
>why that blog post doesn't exist anymore, i don't know.

that blog was moved to a different address, and is still up. Sorry for the misinformation. It wasn't accessible from my side because of my VPN connection perhaps.

https://statusgator.com/blog/has-github-been-down-more-since-its-acquisition-by-microsoft/

how did the outage even happen when they have at least 3 AZs for each region?

>aws was down
>my local nas was also down

>>106962236
A bunch of services globally available lived in us-east-1 with no redundancy. A single point of failure.

wait the internet went down yesterday? i honestly didn't even notice. is /g/ full of goyslop consumers?

>>106962801
so you're telling me those huge companies didn't even pay for more than one AZ per server?

>>106962976
Nope. AWS itself was too lazy to add redundancy to those global services living in a single AZ within one single region.