I appreciated the RiskyBiz episode with the Sentinel one guys where they go over all the ways this could have been prevented if they did real testing
Crowdstrike absolutely deserves the shit they’re getting.
Oh god. Sentinel one is horrible. If they’re taking issue with your testing, you’ve really screwed the pooch
Horrible how? I’ve always thought they were pretty solid in the arena.
Their ftrace hooks caused all disk usage to be serialized, making your multi-core processor single-core when doing anything I/O bound
We saw between 500% - 800% increases in build times with their software installed
Well, that’s spectacular. What do you guys use now?
We’re still using them on machines where performance doesn’t matter
On build machines, they’re on a special VLAN and don’t have endpoint protection, but they only download from a protected mirror
Basically they fucked up and don’t like the criticism from other companies/ customers.
It’s not really criticism, it’s competitors claiming they will never fuck up.
Like, if you found mouse in your hamburger at McDonald’s, that’s a massive fuckup. If Burger King then started saying “you’ll never find anything gross in Burger King food!” that would be both crass opportunism and patently false.
It’s reasonable to criticize CrowdStrike. They fucked up huge. The incident was a fuckup, and creating an environment where one incident could cause total widespread failure was a systemic fuckup. And it’s not even their first fuckup, just the most impactful and public.
But also Microsoft fucked up. And the clients, those who put all of their trust into Microsoft and CrowdStrike without regard to testing, backups, or redundancy, they fucked up, too. Delta shut down, cancelling 4,600 flights. American Airlines cancelled 43 flights, 10 of which would have been cancelled even without the outage.
Like, imagine if some diners at McDonald’s connected their mouths to a chute that delivers pre-chewed food sight-unseen into their gullets, and then got mad when they fell ill from eating a mouse. Don’t do that, not at any restaurant.
Resiliency and security have a lot of layers. The crowd strike bungle was very bad but more than anything it shined a bright spot light on the fact that certain organizations IT orgs are just a house of cards waiting to get blown away.
I’m looking at Delta in particular. Airlines are a critical transportation service and to have issues with one software vendor bring your entire company screeching to a halt is nothing short of embarrassing.
If I were on the board, my first question would be, “where’s our DRP and why was this situation not accounted for?”
House of cards is exactly right. At every IT job I’ve worked, the bosses want to check the DRP box as long as it costs as close to zero dollars as possible, and a day or two of 1-2 people writing it up. I do my best to cover my own ass, and regularly do actual restores, limit potential blast radii, and so on. But at a high level, bosses don’t give AF about defense, they are always on offense (i.e. make more money faster).
This is the first time I’ve heard someone call it a house of cards and I think that fits it perfectly!
Even if that’s the case, how is it Crowdstrike’s place to call these other companies out for claiming something similar will never happen to them? Thus far, it had only ever happened to CS.
It feels like a pattern though. I’ve not seen too much from them but they seem to be saying factually correct stuff. But neither worded correctly nor at the right time.
No, we had Sentinelone take down our company a few months ago. Granted, not a global outage, but it’s something similar. I’m sure that if you went back in news archives, you’d find articles about major Sentinelone outages. I think Crowdstrike is just the biggest one in recent history. It’s certainly not unprecedented.
Even if that’s the case, how is it Crowdstrike’s place to call these other companies out for claiming something similar will never happen to them?
I agree completely, which is why I added that last sentence in an edit. This is a bad look for CrowdStrike, even if I agree with the sentiment.
Thus far, it had only ever happened to CS.
Everybody fucks up now and then. That’s my point. It’s why you shouldn’t trust one company to automatically push security updates to critical production servers without either a testing environment or disaster recovery procedures in place.
I doubt you’ll find any software company, or any company in any industry, that has not fucked up something really important. That’s the nature of commerce. It’s why many security protocols exist in the first place. If everyone could be trusted to do their jobs right 100% of the time, you would only need to worry about malicious attacks which make up only a small fraction of security incidents.
The difference here is that CrowdStrike sold a bunch of clients on the idea that they could be trusted to push security updates to production servers without trsting environments. I doubt they told Delta that they didn’t need DRP or any redundancy, but either way, the failure was amplified by a collective technical debt that corporations have been building into their budget sheets to pad their stock prices.
By all means, switch from CrowdStrike to a competitor. Or sue them for the loss of value resulting in their fuckup. Sort that out in the contracts and courts, because that’s not my area. But we should all recognize that the lesson learned is not to switch to another threat prevention software company that won’t fuck up. Such a company does not exist.
If you stub your toe, you don’t start walking on your hands. You move the damn coffee table out of the pathway and watch where you’re walking. The lesson is to invest in your infrastructure, build in redundancy, and protect your critical systems from shit like this.
you’ll never find anything gross in Burger King food!
Number fifteen…
That’s the first thing I heard in my head lmao
In what way did Microsoft fuck up? They don’t control Crowdstrike updates. Short of the OS files being immutable it seems unlikely they can stop things like this.
Microsoft gave CrowdStrike unfettered access to push an update that can BSOD every Windows machine without a bypass or failsafe in place. That turned out to be a bad idea.
CrowdStrike pushed an errant update. Microsoft allowed a single errant update to cause an unrecoverable boot loop. CrowdStrike is the market leader in their sector and brings in hundreds of millions of dollars every year, but Microsoft is older than the internet and creates hundreds of billions of dollars. CrowdStrike was the primary cause, but Microsoft enabled the meltdown.
Microsoft did not “give Crowdstrike access to push updates”. The IT departments of the companies did.
The security features that Crowdstrike has forces them to run in kernel-space, which means that they will have code running that can crash the OS. They crashed Debian in an almost identical way (forced boot loop) about a month before they did the same to Windows.
Yes, there are ways that Microsoft could rewrite the Windows kernel architecture to make it resistant to this type of failure. But I don’t think there are very many other commercial OS’s that could stop this from happening.
You’re absolutely right, here is an in-depth explanation from Dave Plummer, the guy who wrote the task manager: https://youtu.be/ZHrayP-Y71Q
Microsoft gave CrowdStrike unfettered access to push an update that can BSOD every Windows machine without a bypass or failsafe in place. That turned out to be a bad idea.
They have to give that access by EU ruling:
Microsoft software licensing expert Rich Gibbons said: “Microsoft has received some criticism for the fact that a third party was able to affect Windows at such a deep technical level. It’s interesting that Microsoft has pointed out the fact this stems from a 2009 EU anti-competition ruling that means Microsoft must give other security companies the same access to the Windows kernel as they have themselves.”
Well there’s a provocative anecdote if I’ve ever seen one. Well done.
It’s not really criticism, it’s competitors claiming they will never fuck up.
Not in all cases [podcast warning], sometimes it’s just them pointing out they’re doing silly things like how they test every update and don’t let it out the door with <98% positive returns or having actual deployment rings instead of of yeeting an update to millions systems in less than an hour.
It’s reasonable to criticize CrowdStrike. They fucked up huge. The incident was a fuckup, and creating an environment where one incident could cause total widespread failure was a systemic fuckup. And it’s not even their first fuckup, just the most impactful and public.
Clownstrike deserves every bit of shit they’re getting, and it amazes me that people are buying the bullshit they’re selling. They had no real testing or quality control in place, because if that update had touched test windows boxes it would have tipped them over and they’d have actually known about it ahead of time. Fucking up is fine, we all do it. But when your core practices are that slap dash, bitching about criticism just brings more attention to how badly your processes are designed.
But also Microsoft fucked up.
How did Microsoft fuck up? Giving a security vender kernel access? Like they’re obligated to from previous lawsuits?
And the clients, those who put all of their trust into Microsoft and CrowdStrike without regard to testing, backups, or redundancy, they fucked up, too
Customers can’t test clownstrike updates ahead of time or in a nonprod environment, because clownstrike knows best lol.
Redundancy is not relevant here because what company is going to use different IDR products for primary and secondary tech stacks?
Backups are also not relevant (mostly) because it’s quicker to remediate the problem than restore from backup (unless you had super regular DR snaps and enough resolution to roll back from before the problem.
IMO, clownstrike is the issue, and customers have only the slightest blame for using clownstrike and for not spending extra money on a second IDR on redundant stacks.
Classic corporate “leadership”
We didn’t do nuffin shareholders, please pay
That one for sure…or: Here’s a shiny new thing, customers. You didn’t ask for it, and it actually makes our product a little bit worse…but it’s new. You don’t want to miss out on the new thing, do you?
Well companies serve humans. Be better
I don’t see the other companies fucking up so badly though.
That one recently with the 2.3 billion record data breach is pretty bad, and we collectively had no way to prevent it since it was through a private company.
You don’t? Shit I see it all the time.
well, no, not at the level of taking down half the air traffic, police, ems, fire services, in the free world, you don’t
Crowdstrike is far from the first company to ship bsod and bootloop bugs, they just happened to have the one that had the widest impact. Windows and it’s ecosystem is buggy shit.
You see CrowdStrike level of bad all the time? Where? What? Who?
Equifax
Ok, and what else? I mean saying it happens all the time and listing one doesn’t really prove the point.
That was super bad, and bad things like this do happen. Just saying it’s all the time seems like it’s really hyperbolic.
damn i haven’t used Windows in over a decade. are y’all ok?
Windows users are never OK.
Have you tried turning them off, then turning them on again?
Rofl, like Unix OSes never have problems. Even developers, who are among the most tech savvy users, tend to drag their feet on installing updates unless forced.
Even developers, who are among the most tech savvy users
Doubt
As a sysad, I’ll sign onto this.
i was mostly making a joke about how this absolutely is not a common problem on any platform, not to this degree. and at least when my Arch and Nix systems go down i don’t have anyone to blame but myself. sure, systems have update issues, but a kernel level meltdown that requires a safe mode rescue? that’s literally never happened to me unless it was my fault
I wonder if they’ll end up doing a rename / rebrand if “ClownStrike” continues to haunt them (as it should).
If they do, I’m sure the new name will be some focus tested aberration they pay way too much for.
Yes 🤡
Please proper nomanclature, clownstrike as in 🤡🤽♀️
They deserve to go bankrupt after that level of damage. I won’t be surprised if a class action comes against them for gross negligence within the next few years. They’re cooked and they know it.
Did they try offering a $10 gift card to the other companies? “hah psyche!”
Companies all over the world shutdown because of their incompetence. They do not deserve to be in the security business.
In similar news, Enron says that people should stop being ambulance chasers by calling them out for corruption, fraud, and illegal activities.
ClownStrike had a massive, glaring issue with their main functionality that is THEIR COMPANY’S ONLY REASON FOR EXISTING that has been correctly attributed to criminally inept architecture decisions, no redundancy, no checks, no safety measures, and no accountability.
If I made the executive decision to design a system without any safety measures that could potentially push unchecked, malicious code to 90% of the computers that the business world runs on, I would be sued into dust. Honestly, if there were any justice in the world, the people at CrowdStrike that designed such a shite system should do actual prison time for their ineptitude.
If I made the executive decision to design a system without any safety measures that could potentially push unchecked, malicious code to 90% of the computers that the business world runs on, I would be sued into dust.
Or made a MS CEO…
/me glares at steve balmer
Worry about the future of your stupid failure of a company, not your stupid feelings. Think of the feelings you’ve hurt over your stupid incompetence.
Cry me a half billion dollar river, maybe we can use that money to fix all the damamges it did.
No, its good. They sent out $10 gift cards.
Boo hoo, go fuck yourself.
I am not in the knows about IT security at all, but isn’t fucking up part of any security company? You can make shitty comments all you want but who says your company isn’t the next one to fuck up?
Not wrong, but they fucked up due to incompetence, not just some random preventable accident.
From the technical details I’ve seen, just having a basic testing process/environment should have easily prevented this. That should be the bare minimum.
They whine about this now, after they removed their “shady commentary” towards Microsoft from their website
