[HN Gopher] The Cloudflare outage might be a good thing
       ___________________________________________________________________
        
       The Cloudflare outage might be a good thing
        
       Author : radeeyate
       Score  : 234 points
       Date   : 2025-11-24 03:04 UTC (19 hours ago)
        
 (HTM) web link (gist.github.com)
 (TXT) w3m dump (gist.github.com)
        
       | charcircuit wrote:
       | >It's ironic because the internet was actually designed for
       | decentralisation, a system that governments could use to
       | coordinate their response in the event of nuclear war
       | 
       | This is not true. The internet was never designed to withstand
       | nuclear war.
        
         | chasing0entropy wrote:
         | Arpanet absolutely was designed to be a physically resilient
         | network which could survive the loss of multiple physical
         | switch locations.
        
         | anonym29 wrote:
         | ARPANET was literally invented during the cold war for the
         | specific and explicit purpose of networked communications
         | resilience for government and military in the event major
         | networking hubs went offline due to one or more successful
         | nuclear attacks against the United States
        
           | charcircuit wrote:
           | It literally wasn't. It's an urban myth.
           | 
           | >Bob Taylor initiated the ARPANET project in 1966 to enable
           | resource sharing between remote computers.
           | 
           | >The ARPANET was not started to create a Command and Control
           | System that would survive a nuclear attack, as many now
           | claim.
           | 
           | https://en.wikipedia.org/wiki/ARPANET
        
             | anonym29 wrote:
             | The stated research goals are not necessarily the same as
             | the strategic funding motivations. The DoD clearly
             | recognized packet-switching's survivability and dynamic
             | routing potential when the US Air Force funded the
             | invention of networked packet switching by Paul Baran six
             | years earlier, in 1960, for which the explicit purpose
             | _was_ "nuclear-survivable military communications".
             | 
             | There is zero reason to believe ARPA would've funded the
             | work were it not for internal military recognition of the
             | utility of the underlying technology.
             | 
             | To assume that the project lead was told EVERY motivation
             | of the top secret military intelligence committee that was
             | responsible for 100% of the funding of the project takes
             | either a special kind of naivete or complete ignorance of
             | compartmentalization practices within military R&D and
             | procurement practices.
             | 
             | ARPANET would never have been were it not for ARPA funding,
             | and ARPA never would've funded it were it not for the
             | existence of packet-switched networking, which itself was
             | invented and funded, again, six years before Bob Taylor
             | even entered the picture, for the SOLE purpose of "nuclear-
             | survivable military communications".
             | 
             | Consider the following sequence of events:
             | 
             | 1. US Air Force desires nuclear-survivable military
             | communications, funds Paul Baran's research at RAND
             | 
             | 2. Baran proves packet-switching is conceptually viable for
             | nuclear-survivable communications
             | 
             | 3. His specific implementation doesn't meet rigorous Air
             | Force deployment standards (their implementation partner,
             | AT&T, refuses - which is entirely expectable for what was
             | then a complex new technology that not a single AT&T
             | engineer understood or had ever interacted with during the
             | course of their education), but the concept is now proven
             | and documented
             | 
             | 4. ARPA sees the strategic potential of packet-switched
             | networks for the explicit and sole purpose of nuclear-
             | survivable communications, and decides to fund a more
             | robust development effort
             | 
             | 5. They use academic resource-sharing as the
             | development/testing environment (lower stakes, work out the
             | kinks, get future engineers conceptually familiar with the
             | underlying technology paradigms)
             | 
             | 6. Researchers, including Bob Taylor, genuinely focus on
             | resource sharing because that's what they're told their
             | actual job is, even though that's not actually the true
             | purpose of their work
             | 
             | 7. Once mature, the technology gets deployed for it's
             | originally-intended strategic purposes (MILNET split-off in
             | 1983)
             | 
             | Under this timeline, the sole true reason for ARPA's
             | funding of ARPANET is nuclear-survivable military
             | communication, Bob Taylor, being the military's R&D pawn,
             | is never told that (standard compartmentalization
             | practice). Bob Taylor can credibly and honestly state that
             | he was tasked with implementing resource sharing across
             | academic networks, which is true, but was never the actual
             | underlying motivation to fund his research.
             | 
             | ...and the myth of "ARPANET wasn't created for nuclear
             | survivability" is born.
        
             | oidar wrote:
             | Per interviews, the initial impetus wasn't to withstand a
             | nuclear attack - but after it was first set up, it most
             | certainly a major part of the thought process in design. ht
             | tps://web.archive.org/web/20151104224529/https://www.wired.
             | ..
        
               | charcircuit wrote:
               | >but after it was first set up
               | 
               | Your link is talking about work Baran did before ARPANET
               | was created. The timeline doesn't back your point. And
               | when ARPANET was created after Baran's work with Rand:
               | 
               | >Wired: The myth of the Arpanet - which still persists -
               | is that it was developed to withstand nuclear strikes.
               | That's wrong, isn't it?
               | 
               | >Paul Baran: Yes. Bob Taylor1 had a couple of computer
               | terminals speaking to different machines, and his idea
               | was to have some way of having a terminal speak to any of
               | them and have a network. That's really the origin of the
               | Arpanet. The method used to connect things together was
               | an open issue for a time.
        
               | oidar wrote:
               | Read the whole article. And peruse the oral history here:
               | https://ethw.org/Oral-History:Paul_Baran - the genesis
               | was most definitely related to the cold war.
               | 
               | "A preferred alternative would be to have the ability to
               | withstand a first strike and the capability of returning
               | the damage in kind. This reduces the overwhelming
               | advantage by a first strike, and allows much tighter
               | control over nuclear weapons. This is sometimes called
               | Second Strike Capability."
        
         | bblb wrote:
         | Perhaps. Perhaps not. But it will survive it. It will survive a
         | complete nuclear winter. It's too useful to die, and will be
         | one the first things to be fixed after global annihilation.
         | 
         | But Internet is not hosting companies or cloud providers.
         | Internet does not care if they don't build their systems
         | resilient enough and let the SPOFs creep up. Internet does it's
         | thing and the packets keep flowing. Maybe BGP and DNS could use
         | some additional armoring but there are ways around both of them
         | in case of actual emergency.
        
       | chasing0entropy wrote:
       | Spot on article, but without a call to action. What can we do to
       | combat the migration of society to a centralized corpro-
       | government intertwined entity with no regard for unprofitable
       | privacy or individualism?
        
         | DANmode wrote:
         | Learn how to host anything, today.
        
           | imsurajkadam wrote:
           | Even if you learn to Host, there are many other services that
           | are going to get relied on those centralised platforms, so if
           | you are thinking to Host, every single thing on your own,
           | then it is going to be more work than you can even imagine
           | and definitely super hard to organise as well
        
             | DANmode wrote:
             | Anything.
        
           | randallsquared wrote:
           | Have you tried that? I gave up on hosting my own email server
           | seven or eight years ago, after it became clear that there
           | would be an endless fight with various entities to accept my
           | mail. Hosting a webserver without the expectation that you'll
           | need some high powered DDOS defense seems naive, in the
           | current day, and good luck doing that with a server or two.
        
             | IgorPartola wrote:
             | I have never hosted my own email. It took me roughly a day
             | to set it up on a vanilla FreeBSD install running on
             | Vultr's free tier plan and it has been running flawlessly
             | for nearly a year. I did not use AI at all, just the
             | FreeBSD, Postfix, and Dovecot's handbooks. I do have a fair
             | bit of Linux admin and development experience but all in
             | all this has been a weirdly painless experience.
             | 
             | If you don't love this approach, Mail-in-a-box works
             | incredibly well even if the author of all the Python code
             | behind it insists on using tabs instead of spaces :)
             | 
             | And you can always grab a really good deal from a small
             | hosting company, likely with decades of experience in what
             | they do, via LowEndBox/LowEndTalk. The deal would likely
             | blow AWS/DO/Vultr/Google Cloud out of the water in terms of
             | value. I have been snagging deals from there for ages and I
             | lost a virtual host twice. Once was a new company that
             | turned out to be shady and another was when I rented a VPS
             | in Cairo and a revolution broke out. They brought
             | everything back up after a couple of months.
             | 
             | For example I just bought a lifetime email hosting system
             | with 250GB of storage, email, video, full office suite,
             | calendar, contacts, and file storage for $75. Configuration
             | here is down to setting the DNS records they give you and
             | adding users. Company behind it has been around for ages
             | and is one of the best regarded in the LET community.
        
               | dmoy wrote:
               | It's not insurmountable to set up initially. And when you
               | get email denied from whatever org (your lawyer, your
               | mom, some random business, whatever), each individual one
               | isn't insurmountable to fix. It does get old after
               | awhile.
               | 
               | It also depends on how much you are emailing, and who. If
               | it's always the same set of known entities, you might be
               | totally fine with self hosting. Someone else who's
               | regularly emailing a lot of new people or businesses
               | might incur a lot of overhead. At least worth more than
               | their time than a fastmail or protonmail subscription or
               | whatever.
        
               | randallsquared wrote:
               | I ran my own mail server from 1998 through 2019, and set
               | up a FreeBSD mail server as one of my first contract jobs
               | in 1998 or 1999. I used Sendmail, Exim, Postfix, and
               | qmail at various times. I switched to mail-in-a-box in
               | 2014, and contributed a few minor fixes, then (which I'd
               | forgotten about until I idly looked to see, just now).
               | 
               | Throughout 20 years of running my own mail server for
               | companies, friends, and myself, the additional effort to
               | get commercially-run mail servers to accept mail was both
               | annoying and random ("oh, look, hosted Outlook has
               | started rejecting our mail again..."), and sometimes they
               | don't even send a standard response but just "accept" and
               | blackhole the email. Eventually you find out that someone
               | else in the /24 you're in at Rackspace or DigitalOcean is
               | happily running an open relay, and that's why your IP is
               | having problems. Or any of a dozen similar things.
               | 
               | In 2019, having gotten very tired of this, I gave up and
               | moved my mail handling to Amazon Workmail and SMS, and
               | after setting it up properly once, it's been trouble-free
               | and maintenance-free for half a decade. Compared to some
               | solutions, it's expensive, but not in absolute terms.
        
           | rurban wrote:
           | If you host you are running on my cPanel SW. 70% of the
           | internet is doing that. Also a kinda centralized point of
           | failure, but I didn't hear of any bugs in the last 14 years.
        
         | card_zero wrote:
         | We could quibble about the premise.
        
         | adrianN wrote:
         | Individuals are unlikely to be able to do something about the
         | centralization problem except vote for politicians that want to
         | implement countermeasures. I don't know of any politicians
         | (with a chance to win anything) that have that on their agenda.
        
           | turtletontine wrote:
           | That's called antitrust, and is absolutely a cause you can
           | vote for. Some of the Biden administration's biggest
           | achievements were in antitrust, and the head of the FTC for
           | Biden has joined Mamdani's transition team.
        
           | teiferer wrote:
           | There is a crucial step between having an opinion and voting.
           | It's conversations within society. That's what makes
           | democracy and facilitates change. If you only take your
           | opiniom, isolated from everybody else, and vote from that,
           | there isn't much democracy going on and your chance for
           | change is slim. It's when there is broad conversations
           | happening when movements have an impact.
           | 
           | And that step is here on HN. That's why it's very relevant to
           | observe that that HN crowd is increasingly happy to support a
           | non-free internet. Be it walled gardens, geofencing, etc.
        
       | theideaofcoffee wrote:
       | > They [outages] can force redundancy and resilience into
       | systems.
       | 
       | They won't until either the monetary pain of outages becomes
       | greater than the inefficiency of holding on to more systems to
       | support that redundancy, or, government steps in with clear
       | regulation forcing their hand. And I'm not sure about the latter.
       | So I'm not holding my breath about anything changing. It will
       | continue to be a circus of doing everything on a shoestring
       | because line must go up every quarter or a shareholder doesn't
       | keep their wings.
        
         | morshu9001 wrote:
         | That's ok though, not every website needs 5 9s
        
       | krick wrote:
       | It would be a good thing, if it would cause anything to change.
       | It obviously won't. As if a single person reading this post
       | wasn't aware that the Internet is centralized, and couldn't name
       | specifically a few sources of centralization (Cloudflare, AWS,
       | Gmail, Github). As if it's the first time this happens. As if
       | after the last time AWS failed (or the one before that, or one
       | before...) anybody stopped using AWS. As if anybody _could_
       | viably stop using them.
        
         | captainkrtek wrote:
         | > It would be a good thing, if it would cause anything to
         | change. It obviously won't.
         | 
         | I agree wholeheartedly. The only change is internal to these
         | organizations (eg: CloudFlare, AWS) Improvements will be made
         | to the relevant systems, and some teams internally will also
         | audit for similar behavior, add tests, and fix some bugs.
         | 
         | However, nothing external will change. The cycle of pretending
         | like you are going to implement multi-region fades after a
         | week. And each company goes on continuing to leverage all these
         | services to the Nth degree, waiting for the next outage.
         | 
         | Not advocating that organizations _should /could_ do much, it's
         | all pros/cons. But the collective blast radius is still
         | impressive.
        
           | chii wrote:
           | the root cause is customers refusing to punish these
           | downtime.
           | 
           | Checkout how hard customers punish blackouts from the grid -
           | both via wallet, but also via voting/gov't. It's why they are
           | now more reliable.
           | 
           | So unless the backbone infrastructure gets the same flak,
           | nothing is going to change. After all, any change is
           | expensive, and the cost of that change needs to be worth it.
        
             | MikeNotThePope wrote:
             | Is a little downtime such a bad thing? Trying to avoid some
             | bumps and bruises in your business has diminishing returns.
        
               | aaron_m04 wrote:
               | Depends on the business.
        
               | krige wrote:
               | What's "a little downtime" to you might be work ruined
               | and day wasted for someone else.
        
               | fragmede wrote:
               | It's 2025. That downtime could be be difference between
               | my cat pics not loading fast enough, or someone's
               | teleoperated robot surgeon glitching out.
        
               | bloppe wrote:
               | I remember a Google cloud outage years ago that happened
               | to coincide with one of our customers' massively
               | expensive TV ads. All the people who normally would've
               | gone straight to their website instead got 502. Probably
               | a 1M+ loss for them all things considered.
               | 
               | We got an _extremely_ angry email about it.
        
               | cactusplant7374 wrote:
               | I have a lot of bad days every year. More than I can
               | count. It's just part of living.
        
               | Xelbair wrote:
               | Even more so when most of the internet is also down.
               | 
               | What are customers going to do? Go to competitor that's
               | also down?
               | 
               | It is extremely annoying, will ruin your day, but as
               | movie quote goes - if everyone is special, no one is.
        
               | immibis wrote:
               | They could go to your competitor that's up. If you choose
               | to be up, your competitor's customers could go to you.
        
               | dewey wrote:
               | If it's that easy to get the exact same service / product
               | as another vendor the maybe your competitive advantage
               | isn't so high. If Amazon would be down I'd just wait a
               | few hours as I don't want to sign up on another site.
        
               | MikeNotThePope wrote:
               | I agree. These days it seems like everything is a micro-
               | optimization to squeeze out a little extra revenue.
               | Eventually most companies lose sight of the need to offer
               | a compelling product that people would be willing to wait
               | for.
        
               | throwaway0352 wrote:
               | I think you're viewing the issue from an office worker's
               | perspective. For us, downtime might just mean heading to
               | the coffee machine and taking a break.
               | 
               | But if a restaurant loses access to its POS system (which
               | has happened), or you're unable to purchase a train
               | ticket, the consequences are very real. Outages like
               | these have tangible impacts on everyday life. That's why
               | there's definitely room for competitors who can offer
               | reliable backup strategies to keep services running.
        
               | mallets wrote:
               | Those are examples where they shouldn't be using public
               | cloud in the first place. Should build those services to
               | be local-first.
               | 
               | Using a different, smaller cloud provider doesn't improve
               | reliability (likely makes it worse) if the architecture
               | itself wrong.
        
               | wongarsu wrote:
               | Do any of those competitors actually have meaningfully
               | better uptime?
               | 
               | From a societal level, having everything shut down at
               | once is an issue. But if you only have one POS system
               | targeting only one backend URL (and that backend has to
               | be online for the POS to work) then cloudflare seems like
               | one of the best choices
               | 
               | If the uptime provided by cloudflare isn't enough then
               | the solution isn't a cloudflare competitor, it's the
               | ability to operate offline (which many POS have,
               | including for card purchases) or at least multiple
               | backends with different DNS, CDN, server location etc.
        
             | mopsi wrote:
             | Downtimes happen one way or another. The upside of using
             | Cloudflare is that bringing things back online is their
             | problem and not mine like when I self-host. :]
             | 
             | Their infrastructure went down for a pretty good reason
             | (let the one who has never caused that kind of error cast
             | the first stone) and was brought back within a reasonable
             | time.
        
             | whatevaa wrote:
             | Grid reliability depends on where you live. In some places,
             | UPS or even a generator is a must have. So it's a bad
             | example, I would say.
        
             | LoganDark wrote:
             | > Checkout how hard customers punish blackouts from the
             | grid - both via wallet, but also via voting/gov't.
             | 
             | What? Since when has anyone ever been free to just up and
             | stop paying for power from the grid? Are you going to pay
             | $10,000 - $100,000 to have another power company install
             | lines? Do you even have another power company in the area?
             | State? Country? Do you even have permission for that to
             | happen near your building? Any building?
             | 
             | The same is true for internet service, although personally
             | I'd gladly pay $10,000 - $100,000 to have literally
             | anything else at my location, but there are no proper other
             | wired providers and I'll die before I ever install any sort
             | of cellular router. Also this is a rented apartment so I'm
             | fucked even if there were competition, although I plan to
             | buy a house in a year or two.
        
               | heartbreak wrote:
               | The hyperscalers definitely vote with their wallets.
        
             | tjwebbnorfolk wrote:
             | > the root cause is customers refusing to punish these
             | downtime.
             | 
             | ok how do I punish cloudflare -- build my own globally-
             | distributed content-delivery network just for myself so
             | that I can be "decentralized"?
             | 
             | Or should I go to one of their even-larger competitors like
             | AWS or GCP?
             | 
             | What exactly do you propose?
        
         | ehhthing wrote:
         | With the rise in unfriendly bots on the internet as well as
         | DDoS botnets reaching 15 Tbps, I don't think many people have
         | much of a choice.
        
           | swiftcoder wrote:
           | The cynic in me wonders how much blame the world's leading
           | vendor of DDoS prevention might share in the creation of that
           | particularly problem
        
             | immibis wrote:
             | They provide free services to DDoS-for-hire services and do
             | not terminate the services when reported.
        
               | zamadatix wrote:
               | Not that I doubt examples exist (I've yet to be at a
               | large place with 0 failures on responding to such issues
               | over the years), but it'd be nice if you'd share the
               | specific examples you have in mind if you're going to
               | bother commenting about it. It helps people understand
               | how much is a systemic problem to be interested in vs
               | having a comment which more easily falls into many other
               | buckets instead. I'd try to build trust off the user
               | profile as well, but it proclaims you're shadowbanned for
               | two different reasons - despite me seeing your comment.
               | 
               | One related topic I've seen brought up is Workers abuse
               | https://www.fortra.com/blog/cloudflare-pages-workers-
               | domains..., but that goes against this claim they do
               | nothing when reported.
        
         | sjamaan wrote:
         | Same with the big Crowdstrike fail of 2024. Especially when
         | everyone kept repeating the laughable statement that these guys
         | have their shit in order, so it couldn't possibly be a simple
         | fuckup on their end. Guess what, they don't, and it was. And
         | nobody has realized the importance of diversity for resilience,
         | so all the major stuff is still running on Windows and using
         | Crowdstrike.
        
           | c0l0 wrote:
           | I wrote https://johannes.truschnigg.info/writing/2024-07-impe
           | nding_g... in response to the CrowdStrike fallout, and was
           | tempted to repost it for the recent CloudFlare whoopsie. It's
           | just too bad that publishing rants won't change the darned
           | status quo! :')
        
             | graemep wrote:
             | People will not do anything until something really
             | disastrous happens. Even afterwards memories can fade.
             | Cloudstrike has not lost many customers.
             | 
             | Covid is a good parallel. A pandemic was always possible,
             | there is always a reasonable chance of one over the course
             | of decades. However people did not take it seriously until
             | it actually happened.
             | 
             | A lot of Asian countries are a lot better prepared for a
             | tsunami then they were before 2004.
             | 
             | The UK was supposed to have emergency plans for a pandemic,
             | but it was for a flu variant, and I suspect even those
             | plans were under-resourced and not fit for purpose. We are
             | supposed to have plans for a solar storm but when another
             | Carrington even occurs I very much doubt we will deal with
             | it smoothly.
        
         | testdelacc1 wrote:
         | If anything, centralisation shields companies using a
         | hyperscaler from criticism. You'll see downtime no matter where
         | you host. If you self host and go down for a few hours,
         | customers blame you. If you host on AWS and "the internet goes
         | down", then customers treat it akin to an act of God, like a
         | natural disaster that affects everyone.
         | 
         | It's not great being down for hours, but that will happen
         | regardless. Most companies prefer the option that helps them
         | avoid the ire of their customers.
         | 
         | Where it's a bigger problem is when a critical industry like
         | retail banking in a country all choose AWS. When AWS goes down
         | all citizens lose access to their money. They can't pay for
         | groceries or transport. They're stranded and starving, life
         | grinds to a halt. But even then, this is not the bank's problem
         | because they're not doing worse than their competitors. It's
         | something for the banking regulator and government to worry
         | about. I'm not saying the bank shouldn't worry about it, I'm
         | saying in practice they don't worry about it unless the
         | regulator makes them worry.
         | 
         | I completely empathise with people frustrated with this status
         | quo. It's not great that we've normalised a few large outages a
         | year. But for most companies, this is the rational thing to do.
         | And barring a few critical industries like banking, it's also
         | rational for governments to not intervene.
        
           | DeathArrow wrote:
           | >If anything, centralisation shields companies using a
           | hyperscaler from criticism. You'll see downtime no matter
           | where you host. If you self host and go down for a few hours,
           | customers blame you.
           | 
           | What if you host on AWS and only you go down? How does
           | hosting on AWS shield you from criticism?
        
             | testdelacc1 wrote:
             | This discussion is assuming that the outage is entirely out
             | of your control because the underlying datacenter you
             | relied on went down.
             | 
             | Outages because of bad code do happen and the criticism is
             | fully on the company. They can be mitigated by better
             | testing and quick rollbacks, which is good. But outages at
             | the datacenter level - nothing you can do about that. You
             | just wait until the datacenter is fixed.
             | 
             | This discussion started because companies are actually fine
             | with this state of affairs. They are risking major outages
             | but so are all their competitors so it's fine actually. The
             | juice isn't worth the squeeze to them, unless an external
             | entity like the banking regulator makes them care.
        
           | graemep wrote:
           | > If anything, centralisation shields companies using a
           | hyperscaler from criticism. You'll see downtime no matter
           | where you host. If you self host and go down for a few hours,
           | customers blame you.
           | 
           | Not just customers. Your management take the same view. Using
           | hyperscalers is great CYA. The same for any replacement of
           | internally provided services with external ones from big
           | names.
        
             | testdelacc1 wrote:
             | Exactly. No one got fired for using AWS. Advocating for
             | self-hosting or a smaller provider means you get blamed
             | when the inevitable downtime comes around.
        
           | BlackFly wrote:
           | I think this really depends on your industry.
           | 
           | If you cannot give a patient life saving dialysis because you
           | don't have a backup generator then you are likely facing some
           | liability. If you cannot give a patient life saving dialysis
           | because your scheduling software is down because of a major
           | outage at a third party and you have no local redundancy then
           | you are in a similar situation. Obviously this depends on
           | your jurisdiction and probably we are in different ones, but
           | I feel confident that you want to live in a district where a
           | hospital is reasonably responsible for such foreseeable
           | disasters.
        
             | testdelacc1 wrote:
             | Yeah I mentioned banking because of what I was familiar
             | with but medical industry is going to be similar.
             | 
             | But they do differ - it's never ok for a hospital to be
             | unable to dispense care. But it is somewhat ok for one bank
             | to be down. We just assume that people have at least two
             | bank accounts. The problem the banking regulator faces is
             | that when AWS goes down, all banks go down simultaneously.
             | Not terrible for any individual bank, but catastrophic for
             | the country.
             | 
             | And now you see what a juicy target an AWS DC is for an
             | adversary. They go down on their own now, but surely Russia
             | or others are looking at this and thinking "damn, one
             | missile at the right data Center and life in this country
             | grinds to a halt".
        
         | stingraycharles wrote:
         | It's just a function of costs vs benefits. For most people,
         | building redundancy at this layer costs far too much than the
         | benefits.
         | 
         | If Cloudflare or AWS go down, the outage is usually so big that
         | smaller players have an excuse and people accept that.
         | 
         | It's as simple as that.
         | 
         | "Why isn't your site working?" "Half the internet is down, here
         | read this news article: ..." "Oh, okay, let me know when it's
         | back!"
        
         | tcfhgj wrote:
         | > As if anybody could viably stop using them.
         | 
         | You can, and even save money.
        
         | fragmede wrote:
         | > It obviously won't.
         | 
         | Here's where we separate the men from the boys, the women from
         | the girls, the Enbys from the enbetts, and the SREs from the
         | DevOps. If you went down when Cloudflare went do, do you go
         | multicloud so that can't happen again, or do you shrug your
         | shoulders and say "well, everyone else is down"? Have some
         | pride in your work, do better, be better, and strive for
         | greatness. Have backup plans for your backup plans, and get out
         | of the pit of mediocrity.
         | 
         | Or not, shit's expensive and kubernetes is too complicated and
         | "no one" _needs_ that.
        
           | rkomorn wrote:
           | You make the appropriate cost/benefit decision for your
           | business and ignore apathy on one side and dogma on the
           | other.
        
         | markus_zhang wrote:
         | It's too few and far between. It's gonna make some changes if
         | it's a monthly event. If businesses start to lose connection
         | for 8 hours every month, maybe the bigger ones are going to run
         | for self hosting or at least some capacity of self hosting.
        
           | mkornaukhov wrote:
           | Yeah, agree. But even in case of 8 hour downtime (it's almost
           | 99% SLA) it isn't beneficial for really small firms.
        
         | GuB-42 wrote:
         | Same idea with the Crowdstrike bug, it seems like it didn't
         | have much of on effect on their customers, certainly not with
         | my company at least, and the stock quickly recovered, in fact
         | doing very well. For me, it looks like nothing changed, no
         | lessons learned.
        
           | beanjuiceII wrote:
           | what do you mean no lesson learned? seems like you haven't
           | been paying attention..there's always a lesson learned
        
             | peaseagee wrote:
             | I believe they mean that Crowdstrike learned that they
             | could screw up on this level and keep their customers....
        
               | thewebguyd wrote:
               | That's true of a lot of "Enterprise" software. Microsoft
               | enjoys success from abusing their enterprise customers
               | what seems like daily at this point.
               | 
               | For bigger firms, the reality is that it would probably
               | cost more to switch EDR vendors than the outage itself
               | cost them, and up to that point, CrowdStrike was _the_
               | industry standard and enjoyed a really good track records
               | and reputation.
               | 
               | Depending on the business, there are long term contracts
               | and early termination fees, there's the need to run your
               | new solution along side the old during migration, there's
               | probably years of telemetry and incident data that you
               | need to keep on the old platform, so even if you switch,
               | you're still paying for CrowdStrike for the retention
               | period. It was one (major) issue over 10+ years.
               | 
               | Just like with CloudFlare, the switching costs are higher
               | than outage cost, unless there was a major outage of that
               | scale multiple times per year.
        
               | beanjuiceII wrote:
               | that IS the lesson! there are a million questions i can
               | ask myself about those incidents. What dictates they
               | can't ever screw up? sure it was a big screw up, but
               | understanding the tolerances for screw ups is important
               | to understanding how fast and loose you can play it. AWS
               | has at least a big outage a year, whats the breaking
               | point? risk and reward etc.
               | 
               | I've worked places where every little thing is yak
               | shaved, and places where no one is even sure if the
               | servers are up during working hours. Both jobs paid
               | well.. both jobs had enough happy customers
        
         | ectospheno wrote:
         | I'm pretty cloudflare centric. I didn't start that way. I had
         | services spread out for redundancy. It was a huge pain. Then
         | bots got even more aggressive than usual. I asked why I kept
         | doing this to myself and finally decided my time was worth
         | recapturing.
         | 
         | Did everything become inaccessible the last outage? Yep.
         | Weighed against the time it saves me throughout the year I call
         | it a wash. No plans to move.
        
       | 0x073 wrote:
       | The outage wasn't a good thing, since nothing is changing as a
       | result. (How many outages does cloud flare had?)
        
       | timenotwasted wrote:
       | "Embrace outages, and build redundancy." -- It feels like back in
       | the day this was championed pretty hard especially by places like
       | Netflix (Chaos Monkey) but as downtime has become more expected
       | it seems we are sliding backwards. I have a tendency to rely too
       | much on feelings so I'm sure someone could point me to some data
       | that proves otherwise but for now that's my read on things.
       | Personally, I've been going a lot more in on self-hosting lots of
       | things I used to just mindlessly leave on the cloud.
        
       | stroebs wrote:
       | The problem is far more nuanced than the internet simply becoming
       | too centralised.
       | 
       | I want to host my gas station network's air machine
       | infrastructure, and I only want people in the US to be able to
       | access it. That simple task is literally impossible with what we
       | have allowed the internet to become.
       | 
       | FWIW I love Cloudflare's products and make use of a large amount
       | of them, but I can't advocate for using them in my professional
       | job since we actually require distributed infrastructure that
       | won't fail globally in random ways we can't control.
        
         | Fnoord wrote:
         | Literally impossible? On the contrary; Geofencing is easy. I
         | block all kind of nefarious countries on my firewall, and I
         | don't miss them (no loss not being able to connect to/from a
         | mafia state like Russia). Now, if I were to block FAMAG... or
         | Cloudflare...
        
           | stroebs wrote:
           | Yes, literally impossible. The barrier to entry for anyone on
           | the internet to create a proxy or VPN to bypass your
           | geofencing is significantly lower than your cost to prevent
           | them.
        
             | Aurornis wrote:
             | I don't even understand where this line of reasoning is
             | going. Did you want a separate network blocked off from the
             | world? A ban on VPNs? What are we supposed to believe could
             | have been disallowed to make this happen?
        
             | Dylan16807 wrote:
             | I don't understand why you want to allow any random guy
             | anywhere in the US but not people country hopping on VPNs.
             | For your air machine infrastructure.
             | 
             | It's a bit weird that you can't do this simple thing, but
             | what's the motivation for this simple thing?
        
             | Joel_Mckay wrote:
             | Actually, the 140k Tor exit nodes, VPNs, and compromised
             | proxy servers have been indexed.
             | 
             | It takes 24 minutes to compile these firewall rules, but
             | the black-list along with tripwires have proven effective
             | at banning game cheats. Example, dropping connections from
             | TX with a hop-count and latency significantly different
             | from their peers.
             | 
             | Preemptively banning all bad-reputation cloud IP ranges
             | except whitelisted hosts has zero impact on clients. =3
        
               | Fnoord wrote:
               | I don't have a filter list for compromised proxy servers
               | and VPNs. Do you have a link? I'd be interested in
               | logging such. For Tor, I use [1] (formats in json, txt,
               | md) on OPNsense, but I've also been able to indeed simply
               | parse ASNs (which I currently use for "Twitter, Inc.").
               | 
               | > Preemptively banning all bad-reputation cloud IP ranges
               | except whitelisted hosts has zero impact on clients. =3
               | 
               | This. There's outbound and inbound, and it is very
               | unlikely your print server requires connections from
               | Russia or China (to name an example). You're probably
               | better off making a whitelist, jumphost, or using a VPN
               | with proper authentication to access your services.
               | 
               | Outbound, now that is more difficult to assess. On a
               | desktop, I like a personal firewall for that purpose.
               | Little Snitch on macOS and Open Snitch on Linux have
               | helped me a lot here, but ultimately your hardware
               | firewall is probably lenient on outgoing connections,
               | when you should ask yourself does my network require
               | this, or are they better off with only a HTTP(S) proxy by
               | default?
               | 
               | [1] https://github.com/7c/torfilter
        
         | Aurornis wrote:
         | > and I only want people in the US to be able to access it.
         | That simple task is literally impossible with what we have
         | allowed the internet to become.
         | 
         | Is anyone else as confused as I am about how common anti-
         | openness and anti-freedom comments are becoming on HN? I don't
         | even understand what this comment wants: Banning VPNs? Walling
         | off the rest of the world from US internet? Strict government
         | identity and citizenship verification of people allowed to use
         | the internet?
         | 
         | It's weird to see these comments get traction after growing up
         | in an internet where tech comments were relentlessly pro
         | freedom and openness on the web. Now it seems like every day I
         | open HN and there are calls to lock things down, shut down
         | websites, institute age (and therefore identify) verification
         | requirements. It's all so foreign and it feels like the vibe
         | shift happened overnight.
        
           | dmoy wrote:
           | > Is anyone else as confused as I am about how common anti-
           | openness and anti-freedom comments are becoming on HN?
           | 
           | In this specific case I don't think it's about being anti-
           | open? It's that a business with only physical presence in one
           | country selling a service that is only accessible physically
           | inside the country.... doesn't.... have any need for selling
           | compressed air to someone who isn't like 15 minutes away from
           | one of their gas stations?
           | 
           | If we're being charitable to GP, that's my read at least.
           | 
           | If it was a digital services company, sure. Meatspace in only
           | one region though, is a different thing?
        
             | vpribish wrote:
             | you're being obtuse, GP clearly wants a locked down
             | internet
        
             | teiferer wrote:
             | > In this specific case I don't think it's about being
             | anti-open? It's that a business with only physical presence
             | in one country selling a service that is only accessible
             | physically inside the country.... doesn't.... have any need
             | for selling compressed air to someone who isn't like 15
             | minutes away from one of their gas stations?
             | 
             | But that person might be physically further away at the
             | time they want to order something or gather information
             | etc. Maybe they are on holidays in Spain and want to access
             | their account to pay a bill. Maybe they are in Mexico on a
             | work trip and want to help their aunt back home to use some
             | service for which they need to log in from abroad.
             | 
             | The other day I helped a neighbor (over here in Europe)
             | prepare for a trip to Canada where he wanted to make
             | adjustments to a car sharing account. The website always
             | timed out. It was geofenced. I helped him set up a VPN.
             | That illustrated how locked in this all has become,
             | geofencing without thinking twice.
        
               | dmoy wrote:
               | I guess GP didn't provide enough info, but to me it
               | looked like it was the underlying infra that is networked
               | 
               | That is I'm assuming:
               | 
               | 1. Customers are meatspace only, never use any computer
               | interface 2. The network access is for administration
               | only 3. That administration is exclusively in the US
        
               | Dylan16807 wrote:
               | That's the most obvious answer but if that's the case
               | then restricting to "US" is _way_ too wide in the general
               | case and also too narrow if an employee takes a trip to
               | another country and tries to check in. That simple task
               | is fundamentally flawed to the point it 's not worth
               | worrying about.
        
             | tensegrist wrote:
             | "only need US customers to be able to" vs "want non-US
             | customers to be unable to"
        
             | Aurornis wrote:
             | > In this specific case I don't think it's about being
             | anti-open?
             | 
             | The anti-open part was the mention of "allowed to become",
             | as if we needed to disallow something to achieve this
             | unstated goal.
        
           | thewebguyd wrote:
           | > It's all so foreign and it feels like the vibe shift
           | happened overnight.
           | 
           | The cultural zeitgeist around the internet and technology has
           | changed, unfortunately. But it definitely didn't happen
           | overnight. I've been witnessing it happen slowly over the
           | past 8-10 years, with it accelerating rapidly only in the
           | last 5.
           | 
           | I think it's a combination of special interest groups &
           | nation states running propaganda campaigns, both with bots
           | and real people, and a result of the internet "growing up."
           | Once it became a global, high-stakes platform for finance and
           | commerce, businesses took over, and businesses are
           | historically risk averse. Freedom and openness is no longer a
           | virtue but a liability (for them).
        
         | zrm wrote:
         | > I want to host my gas station network's air machine
         | infrastructure, and I only want people in the US to be able to
         | access it. That simple task is literally impossible with what
         | we have allowed the internet to become.
         | 
         | That task was never simple and is unrelated to Cloudflare or
         | AWS. The internet at a fundamental level only knows where the
         | next hop is, not where the source or destination is. And even
         | if it did, it would only know where the machine is, not where
         | the person writing the code that runs on the machine is.
        
           | teiferer wrote:
           | And that is a good thing and we should embrace it instead of
           | giving in to some idiotic ideas from a non-technical C-suite
           | demanding geofencing.
        
         | asimovDev wrote:
         | not a sysadmin here. why wouldn't this be behind a VPN or some
         | kind of whitelist where only confirmed IPs from the offices /
         | gas stations have access to the infrastructure?
        
           | yardstick wrote:
           | In practice, many gas stations have VPNs to various services,
           | typically via multiple VPN links for redundancy. There's no
           | reason why this couldn't be yet another service going over a
           | VPN.
           | 
           | Gas stations didn't stop selling gas during this outage. They
           | have planned for a high degree of network availability for
           | their core services. My guess is this particular station is
           | an independent or the air pumping solution not on anyone's
           | high risk list.
        
         | Joel_Mckay wrote:
         | Client side SSL certificates with embedded user account
         | identification are trivial, and work well for publicly exposed
         | systems where IPsec or Dynamic frame sizes are problematic
         | (corporate networks often mangle traffic.)
         | 
         | Accordingly, connections from unauthorized users is effectively
         | restricted, but is also not necessarily pigeonholed to a single
         | point of failure.
         | 
         | https://www.rabbitmq.com/docs/ssl
         | 
         | Best of luck =3
        
         | Xelbair wrote:
         | Genuine question - why are you spending time and effort on
         | geofencing when you could spend it on improving your
         | software/service?
         | 
         | It takes time and effort for no gain in any sensible business
         | goal. People outside of US won't need it, bad actors will spoof
         | their location, and it might inconvenience your real customers.
         | 
         | And if you want a secure communication just setup zero-trust
         | network.
        
           | WJW wrote:
           | > bad actors will spoof their location
           | 
           | Isn't that exactly the point? Why are North Korean hackers
           | even allowed to connect to the service, and why is spoofing
           | location still so easy and unverifiable?
           | 
           | Nobody is expected to personally secure their physical
           | location against hostile state actors. My office is not
           | artillery proof, nor does it need to be: hostile actions
           | against it would be an act of war and we have the military to
           | handle those kind of things. But with cybersecurity suddenly
           | everyone is expected to handle everyone from the script
           | kiddie next door to the Mossad. I see the point in OPs post:
           | perhaps it would be good if locking down were a little easier
           | than "just setup zero-trust network".
        
             | Xelbair wrote:
             | you can as easily get attackers from within your own
             | networks, you're falling for fallacy that everything on the
             | 'inside' is secure.
        
               | WJW wrote:
               | Just because one group of attackers is (/might be) inside
               | your network doesn't mean you also have to let all other
               | groups in. There is zero reason to let (say) North
               | Koreans interact with your gas pump API, other than that
               | the internet is set up so that it is virtually impossible
               | to prevent unfriendly parties from contacting your
               | servers.
        
             | Aurornis wrote:
             | > Why are North Korean hackers even allowed to connect to
             | the service,
             | 
             | Asking why some group is "allowed" to use the internet is
             | equivalent to demanding either strict verification or that
             | we cut off some entire country where they reside from the
             | entire internet.
             | 
             | Either that, or someone doesn't understand basic
             | fundamentals of networking and thinks there's some magic
             | solution to this problem.
             | 
             | A common variation of this comment is "why do we allow kids
             | to access <insert topic here>" with demands that something
             | be done about it. Then when something is done about it,
             | there is shock and outrage upon realizing that you can't
             | filter out children without forcing identity verification
             | upon everyone. Similar vibes here, just replace age with
             | demographic.
        
               | WJW wrote:
               | It wouldn't surprise me at all if mandatory online ID
               | verification will become a thing within the next century
               | or so.
        
             | Dylan16807 wrote:
             | North Korea in particular is weird because of sanctions,
             | but pick any country in Europe instead: The user might be a
             | past or future visitor to the gas station and need to
             | access the system even if they're outside the US right now.
             | Or maybe they're actually at the gas station but their
             | phone's data is based in Europe.
             | 
             | Even accurate country tracking is flawed in most
             | situations.
             | 
             | If the goal is specifically "is at the gas station right
             | now" then maybe there's a gap in functionality here, but
             | you could make them connect to the wifi.
             | 
             | Also country-sponsored hackers can easily get a real
             | presence in the US. If country level geoblocking became
             | perfect, they wouldn't be slowed down for more than a week.
        
         | notepad0x90 wrote:
         | Is Cloudflare having more outages than aws, gcp or azure?
         | Honestly curious, I don't know the answer.
        
           | nananana9 wrote:
           | Definitely not.
           | 
           | I was a bit shocked when my mother called me for IT help and
           | sent me a screenshot of a Cloudflare error page with
           | Cloudflare being the broken link and not the server. I
           | assumed it's a bug in the error page and told her that the
           | server is down.
        
       | L-four wrote:
       | It's a tragedy of the commons. Even if you don't use Cloudflare
       | does it matter if no one can pay for your products.
        
       | tonyhart7 wrote:
       | I don't like this argument since you can applied this argument to
       | google,microsot,aws,facebook etc
       | 
       | Tech world is dominated by US company and what is alternative to
       | most of these service???? its a lot fewer than you might think
       | and even then you must make a compromise in certain areas
        
       | throwaway81523 wrote:
       | Now just wait til every country on earth really does replace most
       | of its employees with ChatGPT... and then OpenAI's data center
       | goes offline with a fiber cut or something. All work everywhere
       | stops. Cloudflare outage is nothing compared to that.
        
         | DeathArrow wrote:
         | That's why it's better to have redundancy. Hire Claude and
         | Deepseek, too.
        
         | teiferer wrote:
         | > goes offline with a fiber cut
         | 
         | If a fiber cut brings your network down then you have
         | fundamental network design issues and need to change hiring
         | practices.
        
         | delaminator wrote:
         | That was this outage. ChatGPT and Claude are both behind
         | Clouflare's bot detection. You couldn't log into either Web
         | frontends.
         | 
         | And the error message said you were blocking them. We had
         | support tickets coming in demanding to know why ChatGPT was
         | being blocked.
         | 
         | We also couldn't log into our supplier's B2B system to place
         | our customer orders.
         | 
         | So all the advice of "just self host" is moot when you're in a
         | food web.
        
       | oidar wrote:
       | I wonder what would life without cloudflare look like? What
       | practices would fill the gaps if a company didn't - or wasn't
       | allowed to -- satisfy the the concerns that cloudflare fills.
        
         | immibis wrote:
         | Pretty much exactly like it does now but with less captchas.
         | 
         | Fun fact: Headless browsers can easily pass cloudflare captchas
         | automatically. They're not actually captchaing - they're just a
         | placebo. You just need to be coming from a residential IP
         | address and using a real browser.
        
           | hombre_fatal wrote:
           | > Pretty much exactly like it does now but with less
           | captchas.
           | 
           | This just isn't true. e.g. I saw a 30x increase in traffic on
           | my forum due to AI bots that I had to use CF to block.
           | 
           | CF is mainly empowered by the naive ideals of the internet's
           | design that never built-in countermeasures against bad
           | actors. You're expected to just deal with it yourself
           | somehow. And that means outsourcing it, especially as
           | residential IP address botnets on unlimited ISP data plans
           | become cheaper and cheaper.
           | 
           | Just ask yourself why web hosting providers themselves can't
           | offer services at CF's level. It's because it's too hard of a
           | problem even for them.
        
             | immibis wrote:
             | You didn't _have to_ use CF to block them. You chose to use
             | CF to block them. How was your experience with Anubis or
             | https://git.gammaspectra.live/git/go-away?
             | 
             | Or you could simply... serve the requests. If your normal
             | traffic is only, like, 1 request per minute, then 30x that
             | is still pretty low and there's no actual reason to worry
             | about it.
             | 
             | Web hosting providers don't offer bot blockers because
             | first, they have no reason to care, and second, they can
             | serve the requests, and third, some of them want to upsell
             | you on bandwidth (you should prefer the ones with unmetered
             | bandwidth).
             | 
             | BTW AFAIK there's still zero evidence that the massive DDoS
             | wave has anything at all to do with AI. It could be, say,
             | one of Russia's many small avenues of trying to break the
             | West, or Cloudflare trying to get more business, or the NSA
             | trying to make Cloudflare get more business because it's
             | tapped into Cloudflare.
        
       | zie1ony wrote:
       | My friend wasn't able to do RTG during the outage. They had to
       | use ultrasound machine on his broken arm to see inside.
        
         | Aurornis wrote:
         | > My friend wasn't able to do RTG during the outage.
         | 
         | What is RTG?
        
           | soni96pl wrote:
           | X-ray
        
           | digestives wrote:
           | X-ray, in some languages (like Polish) the abbreviation comes
           | from https://en.wikipedia.org/wiki/Roentgen_(unit)
        
           | teiferer wrote:
           | Wilhelm Rontgen, Nobel Prize in 1901, experimentally
           | discovered X-rays.
        
       | Surac wrote:
       | The thing I learned from the incident is that rust offer a unpack
       | function. It puzzles me why the hell they build such a function
       | in the first place.
        
         | aw1621107 wrote:
         | > It puzzles me why the hell they build such a function in the
         | first place.
         | 
         | One reason is similar to why most programming languages don't
         | return an Option<T> when indexing into an
         | array/vector/list/etc. There are always tradeoffs to make,
         | especially when your strangeness budget is going to other
         | things.
        
       | almosthere wrote:
       | how many people are still on us-east-1
        
         | mcny wrote:
         | My old employer used azure. It irritated me to no end when they
         | said we must rename all our resources to match the convention
         | of naming everything US East as "eu-" because (Eastern United
         | States I guess)
         | 
         | A total clown show
        
       | 0xbadcafebee wrote:
       | Centralization has nothing to do with the problems of society and
       | technology. And if you think the internet is _all_ controlled by
       | just a couple companies, you don 't actually understand how it
       | works. The internet is _wildly_ decentralized. Even Cloudflare
       | is. It offers tons of services, all of which are completely
       | optional and can be used individually. You can also stop using
       | them at any time, and use any of their competitors (of which
       | there are many).
       | 
       | If, on the off chance, people just get "addicted" to Cloudflare,
       | and Cloudflare's now-obviously-terrible engineering causes
       | society to become less reliable, then people will respond to
       | that. Either competitors will pop up, or people will depend on
       | them less, or governments will (finally!) impose some regulations
       | around the operation of technical infrastructure.
       | 
       | We have actually _too much_ freedom on the Internet. Companies
       | are free to build internet systems any way they want - including
       | in very unreliable ways - because we impose no regulations or
       | standards requirements on them. Those people are then free to
       | sell products to real people based on this shoddy design, with no
       | penalty for the products falling apart. So far we haven 't had
       | any gigantic disasters (Great Chicago Fire, Triangle Shirtwaist
       | Factory Fire, MGM Grand Hotel Fire), but we have had major
       | disruptions.
       | 
       | We already dealt with this problem in the rest of society.
       | Buildings have building codes, fire codes, electrical codes. They
       | prescribe and require testing procedures, provide standard
       | building methods to ensure strength in extreme weather, resist a
       | spreading fire long enough to allow people to escape, etc. All
       | measures to ensure the safety and reliability of the things we
       | interact with and depend on. You can build anything you want -
       | say, a preschool? - but you aren't allowed to build it in a
       | shoddy manner. We have that for physical infrastructure; now we
       | need it for virtual infrastructure. A software building code.
        
         | DeathArrow wrote:
         | Centralization means having a single point of failure for
         | everything. If your government, mobile phone or car stops
         | working, it doesn't mean all governments, all cars and all
         | mobile phones stop working.
         | 
         | Centralization makes mass surveillance easier, makes
         | selectively denying of service easier. Centralization also
         | means that once someone hacks into the system, he gains access
         | to all data, not just a part of it.
        
       | vasco wrote:
       | I'll die on the hill that centralization is more efficient than
       | decentralization and that rare outages of hugely centralized
       | systems that are otherwise highly reliable are much better than
       | full decentralization with much worse reliability.
       | 
       | In other words, when AWS or Cloudflare go down it's catastrophic
       | in the sense that everyone sees the issues at the same time, but
       | smaller providers usually have much more ongoing issues, that
       | just happen to be "chronic" vs "acute" pains.
        
         | Xelbair wrote:
         | >I'll die on hill that hyperoptimized systems are more
         | efficient than anti-fragile.
         | 
         | Of course they are, the issue is what level of failure were
         | going to accept.
        
         | GeneralMaximus wrote:
         | Efficient in terms of what, exactly?
         | 
         | There are multiple dimensions to this problem. Putting
         | everything behind Cloudflare might give you better uptime,
         | reliability, performance, etc. but it also has the effect of
         | centralizing power into the hands of a single entity. Instead
         | of twisting the arms of ten different CXOs, your local
         | politician now only needs to twist the arm of a single CXO to
         | knock your entire business off the internet.
         | 
         | I live in India, where the government has always been hostile
         | to the ideals of freedom of speech and expression. Complete
         | internet blackouts are common in several states, and major ISPs
         | block websites without due process or an appeals mechanism.
         | Nobody is safe from this, not even Github[1]. In countries like
         | India, decentralization is a preventative measure.
         | 
         | [1] https://en.wikipedia.org/wiki/Censorship_of_GitHub#India
         | 
         | And I'm not even going to talk about abuse of monopoly power
         | and all that. What happens when Cloudflare has their Apple
         | moment? When they jack up their prices 10x, or refuse to serve
         | customers that might use their CDNs to serve "inappropriate"
         | content? When the definition of "inappropriate" is left fuzzy,
         | so that it applies to everything from CSAM to political
         | commentary?
         | 
         | No thanks.
        
           | vasco wrote:
           | The fix to government censorship must be political, not
           | technical.
        
         | torginus wrote:
         | And the irony is that people are pushing for decentralization
         | like microservices and k8s - on centralized platforms like AWS.
        
       | rzerowan wrote:
       | So were going backwards to a world where there are basically 5
       | computers running everything and everyone is basically accessing
       | the world through a dumb terminal.Even though the digital slab in
       | our pockets has more compute than a roomful of the early gen
       | devices. Hopefully critical infrashifts back to managed metal or
       | private clouds - dont see it though with the last decades of
       | cloud evangalism to move all legacy systems to the cloud doesnt
       | look like reversing anytime soon.
        
         | zwnow wrote:
         | I agree considering all the Cloudflare AWS Azure apologists I
         | see all around... Learning AWS already is the #1 tip on social
         | media to "become employed as a dev in 2025 guaranteed" and I
         | always just sigh when seeing this. I wouldnt touch it with a
         | stick.
        
         | fragmede wrote:
         | Yeah it's crazy to realize it takes a room of electronics for
         | me to get my (g)mail. The more things change, the more they
         | stay the same, eh?
        
       | tomschwiha wrote:
       | For me personally I didn't notice the downtime in the first hour
       | or so. When using some website assets were not loading, but
       | that's it. Turnstile outage maybe impacted me most. Could be
       | because I'm EU based and Cloudflare is not "so" widespread here
       | as in other parts of the world.
        
       | notepad0x90 wrote:
       | Does the author of this post not see the irony of posting this
       | content on Github?
       | 
       | My counter argument is that "centralization" in a technical sense
       | isn't about what company owns things but how services are
       | operated. Cloudflare is very decentralized.
       | 
       | Furthermore, I've seen regional outages caused by things like
       | anchors dropped by ships in the wrong place, a shark eating a
       | cable. Regional power outages caused by squirrels,etc... outages
       | happen.
       | 
       | If everyone ran their own server from their own home, AT&T or
       | Level3 could have an outage and still take out similar swathes of
       | the internet.
       | 
       | With CDNs like cloudflare, if Level3 had an outage, your website
       | won't be down because your home or VPS host's upstream transit
       | happens to be Level3 (or whatever they call themselves these
       | days) because your content (at least static) is cached globally.
       | 
       | The only real reasonable alternative is something like ipfs, web3
       | and similar talk.
       | 
       | Cloudflare has always called itself a content transport provider,
       | think of it as such. But also, Cloudflare is just one player,
       | there are several very big players. Every big cloud provider has
       | a competing product, not to mention companies like Akamai.
       | 
       | People are rage posting about cloudflare, especially because it
       | has made CDNs accessible to everyone. You can easily setup a free
       | cloudflare account and be on your merry way. This isn't something
       | you should be angry about. You're free to pay for any number of
       | other cdns, many do.
       | 
       | If you don't like how Cloudflare has so much market share, then
       | come up with a similarly competitive alternative and profit. Just
       | this HN thread alone is enough for me to think there is a market
       | for more players. Or, just spread the word about the competition
       | that exists today. Use frontdoor, cloudfront, netlify, flycdn,
       | akamai,etc... It's hardly a monopoly.
        
       | jcattle wrote:
       | "The Cloudflare outage was a good thing [...] they're a warning.
       | They can force redundancy and resilience into systems."
       | 
       | - he says. On Github.
        
         | Afforess wrote:
         | Thanks for doing the meme! https://knowyourmeme.com/memes/we-
         | should-improve-society-som...
         | 
         | You are very intelligent!
        
           | jcattle wrote:
           | That's fair. However I don't think I would have wrote that if
           | those thoughts were shared on a blogging platform.
           | 
           | Most blogging platforms do not qualify as critical
           | infrastructure. GitHub with all its CI/CD and supply chain
           | attacks does.
           | 
           | There is a certain particular irony of this being written on
           | critical (centralized) infrastructure without any apparent
           | need.
           | 
           | Maybe it was intended, maybe not, in any case I found it
           | funny.
        
             | rkomorn wrote:
             | I agree. I think the whole point is someone like TFA author
             | has a pretty broad choice of places they can choose to
             | publish this and choosing GitHub is somewhat ironic.
             | 
             | Reminds me of the guy who posted an open letter to Mark
             | Zuckerberg like "we are not for sale" on LinkedIn, a place
             | that literally sells access to its users as their main
             | product.
        
       | nicman23 wrote:
       | i hate that i cannot just scrape things for me usage and i have
       | to use things like camufox instead of curl
        
       | torginus wrote:
       | What happens if you don't use Cloudflare and just host everything
       | on a server?
       | 
       | Can't you run a website like that if you don't host heavy
       | content?
       | 
       | How common are DDOS attacks anyway, and aren't there local (to
       | the server), that analyze user behavior to a decent accuracy (at
       | least it can tell they're using a real browser and behaving more
       | or less like a human would, making attacks expensive).
       | 
       | Can't you buy a list of ISP ranges from a GeoIP provider (you
       | can), at least then you'd know which addresses belong to real
       | humans.
       | 
       | I don't think botnets are that big of a problem (maybe in some
       | obscure places of the world, but you can temp rangeban a certain
       | IP range, if there's a lot of suspicious traffic coming from
       | there).
       | 
       | If lots of legit networks (as in belonging to people who are
       | paying an ISP for their network connections) have botnets, that's
       | means most PCs are compromised, which is a much more severe
       | issue.
        
         | dijit wrote:
         | Yeah, you can.
         | 
         | Lots of people use raspberry pi's for this, which is a smidge
         | anaemic for some decent load (HN Hug Of Death)- even an Intel
         | N100 is more grunt, for context.
         | 
         | This makes people think that their self hosting setup can
         | _never_ handle HN load; because when they see people talking
         | about self hosting the site goes down.
        
           | rainonmoon wrote:
           | Most people shouldn't use a Pi because most people can't
           | configure a web server securely. A VPS would be a better
           | option for just about everybody trying to "self-host" whether
           | they put Cloudflare in front of it or not.
        
             | dijit wrote:
             | in both cases you're setting up a webserver.
             | 
             | I guess you're concerned about lateral network movement?
             | Justified, but as long as it's patched it's going to be
             | just as secure.
        
               | rainonmoon wrote:
               | You're right, but with an asterisk. I don't care if my DO
               | droplet gets popped with an RCE. I do care if someone
               | establishes persistence in my home.
        
               | 1718627440 wrote:
               | You can have different networks in your physical home.
        
               | rainonmoon wrote:
               | And?
        
               | 1718627440 wrote:
               | Meaning your internal network and your publicly hosted
               | services need to not to be in the same network.
        
         | dewey wrote:
         | Botnets use real residential connections not just data centers.
         | So your static list of "real people" doesn't really make a
         | difference.
        
         | justsomehnguy wrote:
         | > What happens if you don't use Cloudflare and just host
         | everything on a server?
         | 
         | It works.
         | 
         | > Can't you run a website like that if you don't host heavy
         | content?
         | 
         | Even with a heavy content - question is how many visitors do
         | you have. If there is one once an hour you would suffice on a
         | 100Mbit/Unlim connection.
         | 
         | > How common are DDOS attacks anyway
         | 
         | Extremely rare. 99% of _sites_ never experience it, 1% do have
         | _some_ trouble because somebody nearby is being DDoS 'ed.
         | 
         | > and aren't there local (to the server), that analyze user
         | behavior to a decent accuracy (at least it can tell they're
         | using a real browser and behaving more or less like a human
         | would, making attacks expensive).
         | 
         | No point, you can't do anything anyway - it's a _denial_ of
         | service so there are gigabytes of trash flowing your way.
         | 
         | > Can't you buy a list of ISP ranges from a GeoIP provider (you
         | can), at least then you'd know which addresses belong to real
         | humans.
         | 
         | No point. If you are not being DDoS'ed then you just spent
         | money and time (ie money) on useless preventive measure you
         | never use. And when (if) it would come you can't do anything
         | anyway, because it's a distributed denial of service attack.
         | 
         | > I don't think botnets are that big of a problem (maybe in
         | some obscure places of the world, but you can temp rangeban a
         | certain IP range, if there's a lot of suspicious traffic coming
         | from there).
         | 
         | It's not a DDoS if you can filter at the endpoint.
        
       | miki123211 wrote:
       | I don't know how many times I need to say this, but I will die on
       | this hill.
       | 
       | Centralized services don't decrease redundancy. They're usually
       | far more redundant than whatever homegrown solution you can come
       | up with.
       | 
       | The difference between centralized and homegrown is mostly
       | psychological. We notice the outages of centralized systems more
       | often, as they affect everything at the same time instead of
       | different systems at different times. This is true even if, in a
       | hypothetical world with no centralization, we'd have more total
       | outage time than we do now.
       | 
       | If your gas station says "closed" due to a problem that only
       | affects their own networks, people usually go "aah they're
       | probably doing repairs or something", and forget about the
       | problem 5 minutes later. If there's a Cloudflare outage...
       | everybody (rightly) blames the Cloudflare outage.
       | 
       | Where this becomes a problem is when correlated failures are
       | actually worse than uncorrelated ones. If Visa goes down, it's
       | better if Mastercard stays up, because many customers have both
       | and can use the other when one doesn't work. In some ways, it's
       | better to have 30 mins of Visa outages today and 30 mins of
       | Mastercard outages tomorrow, than to have just 15 mins of
       | correlated outages in one day.
        
         | dgan wrote:
         | > Centralized services don't decrease redundancy
         | 
         | Alright, but it creates a failure correlation where previously
         | there was none
        
           | silvestrov wrote:
           | Have you ever heard of the "sendmail worm", aka Morris Worm ?
           | 
           | https://en.wikipedia.org/wiki/Morris_worm
           | 
           | You can definitely have failure correlation without having
           | centralized services.
        
         | masfuerte wrote:
         | In my experience services aren't failing due to a lack of
         | redundancy but due to an excess of complexity. With the move to
         | the cloud we are continually increasing both redundancy and
         | complexity and this is making the problem worse.
         | 
         | I have a cheap VPS that has run reliably for a decade except
         | for a planned hour of downtime. Which was in the middle of the
         | night when no-one cared. Amazon is more reliable in theory. My
         | cheap VPS is more reliable in practice.
        
         | lloeki wrote:
         | "redundancy" might not be there correct word. If we had a
         | single worldwide mega-entity serving 100% of the internet it
         | would be both a monopoly and would have tons of redundant
         | infrastructure.
         | 
         | But it would also be quite unified; the system, while full of
         | redundancies, as a whole is a unique one operated the same way
         | end to end; by virtue of it being a single system handled in a
         | uniform way, a single glitch could bring it all down. There is
         | no diversity in the system's implementation, the monoculture
         | itself makes it vulnerable.
        
         | freeplay wrote:
         | The problem is creating a single point of failure.
         | 
         | There's no doubt a VM in AWS is exponentially more redundant
         | than my VM running on a couple of Intel NUCs in my closet.
         | 
         | The difference is, when I have a major outage, my blog goes
         | down.
         | 
         | When EC2 has a major outage, all of the blogs go down. Along
         | with Wikipedia, Starbucks, and half the internet.
         | 
         | That single point of failure is the issue.
        
           | YetAnotherNick wrote:
           | Single point of failure means exactly opposite of what you
           | think it means. If my work depends on 5 services to be up,
           | each service would be a single point of failure, and
           | correlation of failure is good for probability that I can do
           | my work.
        
             | smj-edison wrote:
             | This is a really interesting point, because I could see a
             | situation where your application requires integration with
             | say 10 services. If they all run on AWS, they either all go
             | down or all run together. If they're all self-hosted,
             | there's a good chance that at any time one of the ten is
             | down, and so your service can't run.
        
             | freeplay wrote:
             | I see what you're saying but I have to push back.
             | 
             | "If one thing I need is going to be down, everything might
             | as well be down."
             | 
             | If I have a product with 5 dependencies and one of them is
             | down, there's things I can do to partially mitigate. A
             | circuit breaker would allow my thing to at least stay up
             | and responsive. Maybe I could get a status message up and
             | turn off a feature flag to disable what calls that
             | dependency.
             | 
             | On the other hand, if all my dependencies are down _AND_
             | the management layer is down _AND_ the AWS portal is not
             | functioning correctly, I 'm pretty much SOL.
             | 
             | Massive centralization is never, ever a good thing for
             | anyone other than the ones who are doing the centralizing.
        
       | chrisjj wrote:
       | True title: The Cloudflare outage was a good thing
        
       | joeblubaugh wrote:
       | meta: why are we rewriting such anodyne titles? "was" -> "might
       | be" undermines the author's point
        
       | mrasong wrote:
       | Yeah, when it went down, a bunch of the sites I use every day
       | just stopped working.
       | 
       | That's when I realized it's basically one of the backbone pieces
       | of the entire internet.
        
       | ovo101 wrote:
       | Outages like this highlight just how much of the internet's
       | resilience depends on a single provider. In a way, it's a healthy
       | reminder: if one company's hiccup can take down half the web,
       | maybe we've over-centralized. A "good thing" only if it sparks
       | more serious conversations about redundancy, multi-provider
       | strategies, and reducing monoculture risk. Otherwise, we'll just
       | keep repeating the same failure modes at larger scales.
        
       | rafaelcosta wrote:
       | I don't get why this applies on the Cloudflare outage but not on
       | the AWS ones... I'd argue that the big cloud providers are WAY
       | more impactful when they go down than Cloudflare. The only
       | difference is that the average techie uses Cloudflare more and
       | sees the impact more, but this point was already there before...
        
       | SirMaster wrote:
       | If these systems are as important as they say, it's surprising to
       | me that they are not built with backups and redundancies in place
       | like other mission critical things are engineered and built with.
        
       | tjwebbnorfolk wrote:
       | Every HN comment seems to say the same thing: downtime is
       | inexcusable and the centralization of these services is ruining
       | the internet.
       | 
       | I still don't see the big deal. 12 hours of downtime once every
       | couple years isn't the end of the world. So people can't log into
       | their bank website for a few hours -- banks used to only be open
       | for like 4 hours a day and somehow we all survived. Twitter is
       | down? Oh what a tragedy. Customers get some refunds, Cloudflare
       | fixes the issue, and people move on with life.
       | 
       | Cars still break down occasionally after 100+ years of
       | engineering for reliability and safety. The power still goes out
       | every now and then. Cook on the stove. The cost of making
       | everything perfect all the time just isn't worth it.
       | 
       | I run my own servers on my own network and do not use Cloudflare.
       | My stuff goes down too. And it's "decentralized" in the way you
       | think the internet "should" be, which entails its own risks. So
       | what do you all want, exactly? A public lashing of every
       | developer at Cloudflare who pushes a bug to prod? A congressional
       | investigation? I just don't understand the outrage here.
       | 
       | Stuff breaks occasionally. Get used to it, and design
       | accordingly.
        
         | joshuamcginnis wrote:
         | From a consumers perspective, that makes sense. From a
         | business's perspective, downtime can mean significant loss of
         | revenue or new business opportunity.
        
           | tjwebbnorfolk wrote:
           | The costs of perfection are much, much greater. Are you
           | willing to pay 2-3x the cost of everything to go from 99.999%
           | to 100.0000000% uptime?
           | 
           | Probably the only thing in existence with 100.00% uptime are
           | our nuclear missile command and control systems. Like, even
           | my pen runs out of ink sometimes. It's just crazy how hard it
           | is to have stuff work all the time.
        
           | ahepp wrote:
           | I wonder if consolidation actually makes this less of an
           | issue for businesses?
           | 
           | If my website is down, but my competitors' isn't, I might
           | lose business to them. If my competitor's website is also
           | down, where are the customers gonna go?
        
         | rigrassm wrote:
         | > So people can't log into their bank website for a few hours,
         | banks used to only be open for like 4 hours a day and somehow
         | we all survived.
         | 
         | 1. I believe it's payment processing systems not functioning
         | properly that causes real problems for people and not simply
         | bank websites being down. Especially given...
         | 
         | 2. Banks being closed so much back when cash/checks were
         | actually widely used wasn't an issue because you could just pop
         | over to an ATM or whip out a checkbook. In today's system,
         | every single purchase you make requires communication between
         | the merchant, your bank, and any number of middlemen via the
         | internet.
         | 
         | Yeah, cash is still used today but I've been noticing even
         | things like school sports events have stopped taking cash all
         | together and simply post a QR code to buy from your phone.
         | 
         | That is unless the school has crap cell reception (with no
         | public Wi-Fi either!), Cloudflare shits the bed, Visa thinks
         | you're buying porn, you locked your debit card and now can't
         | unlock it cuz the website is down, or any one of the million
         | things that break all the time. Replace school sports event
         | with literally every single things that requires a financial
         | transaction and it's easy to see how even a short outage can
         | lead to actual harm being realized.
        
       | YmiYugy wrote:
       | It's worth considering the counter factual. Let's say there would
       | be a few dozen semi popular DDoS services. Would that be better?
       | Some assumptions: The services would be slightly less effective
       | and also have worse downtimes. You could argue that Cloudflare is
       | coasting on a monopoly and that competition would drive them to
       | improve, but I'm pretty confident that DDoS protection it one of
       | those things were having a large network to absorb attacks and a
       | large team to monitor them if very valuable. I submit as evidence
       | that Cloudflare has been doing well despite the 3 big cloud
       | providers offering DDoS protection.
       | 
       | So what would be the result of a highly decentralized but
       | slightly worse and less reliable DDoS protection? I'd argue that
       | for a lot of things this wouldn't be an improvement. Cloudflare
       | being so dominant means lot's of things go down simultaneously.
       | But that only matters for fungible services, e.g. if a schools
       | education portal goes down, it doesn't matter if all the other
       | education portals are also down. There are cases where it matters
       | like the tyre pumps. I'd argue that these devices have no reason
       | to be reliant on an online connection to begin with. I think
       | cloud services as a whole have massively improved the reliability
       | of internet services. In almost all cases reducing the overall
       | amount of outages is a higher priority than preventing outage
       | correlations.
        
       ___________________________________________________________________
       (page generated 2025-11-24 23:01 UTC)