Live data from Hacker News

Cloudflare was down

cloudflarestatus.com

311–320 of 499 posts

Re: Cloudflare was down

#311
post #9

Discord is entirely down right now, both the website and the app itself. Amusingly, a lot of the sites that normally track outages are also down, which made me think it was my internet at first. Downdetector, monitortheinternet, etc. Lots of other big sites that are down: Patreon, npmjs, DigitalOcean, Coinbase, Zendesk, Medium, GitLab (502), Fiverr, Upwork, Udemy Edit: 15 min later, looks like things are starting to…

Doordash too. My first order. On my wife's birthday. Ah, well. This too shall pass.

I did a pickup order a few weeks ago when of all things Tmobile SMS went down for 3+ hours. I couldn't go in the restaurant (covid) and I couldn't text them the parking # I was sitting at in a packed parking lot. I got a flood of about 50 texts a few hours later. Sat there for about an hour waiting for a $9 sandwich. I have no idea if they didn't get my order until late, or if they finally realized it was me or what. About 45 minutes in I decided to just give up on the day and take a nap, woke up to a door knock.

Re: Cloudflare was down

#312
Friendly reminder (and notes to myself):

Don't use Namecheap and Cloudflare at the same time.

Namecheap is using cloudflare. So if cloudflare is down, you can't change DNS settings on Namecheap as well!

Re: Cloudflare was down

#313

"All systems operational" What's the point of a status page if it doesn't reflect the real status... It's either the status page goes down with everything else or the status page is wrong. Great. EDIT: Looks like it's accurate now, 20 minutes later.

IBM Cloud status is pretty much always green... although we have issues pretty much every week.

They’re still using Lotus Notes for the tracking.

Re: Cloudflare was down

#314

Earlier quoted context omitted.

https://status.discord.com/ is down. Wow.

It's hosted by statuspage.io, and their own status page was also down (metastatuspage.com). It is now back up, but their page shows the outage.

Status turtles all the way down, it seems!

Re: Cloudflare was down

#315

It really defies the original vision of the internet to have so many services depend on a single company. Almost every news site I was reading dropped off at once. I thought for a second that I lost internet in my own house.

Agreed, but the real problem is DDoS and nobody seems to know how to globally solve it. Fighting DDoS is expensive, so you see consolidation. It's well and good to live in a tiny farming town but when raiders start attacking every week, those castle walls and guards start to look really appealing.

Re: Cloudflare was down

#316
post #194

Our TPU management page is also down: https://www.tensorfork.com/tpus Seems cloudflare took out a good chunk of the internet temporarily. Doesn’t HN use cloudflare? Why did it survive? (I haven’t looked for about a year, but I seem to remember HN being proxied behind CF at one point.)

HN went off Cloudflare a couple years ago. https://news.ycombinator.com/item?id=18188832 https://news.ycombinator.com/item?id=21799045

How do you deal with DDoS attacks?

Re: Cloudflare was down

#317
post #136

We can't keep going on like this. The vulnerability of centralised internet infrastructure is a huge problem for everyone. Somebody, somewhere, really ought to sort it all out

Be the change you want to see in the world :) There are no somebodys somewheres. One question is how to do DDoS protection without somebody like Cloudflare. Some new protocol for edge caching, perhaps?

DDoS has two components:

a) complexity: trick your servers into doing something hard

b) volumetric: overwelm your servers with a lot of traffic

c) volumetric part two: overwelm your servers with a lot of requests, so you respond with a lot of traffic

A and C are things you can work on your self --- try to limit the amount of work your server does in response to requests, and/or make resource consuming responses require resource consuming requests; and monitor and fix hotspots as they're found.

B is tricky, there's two ways to solve volumentric attack; either have enough bandwidth to drop the packets on your end, or convince the other end to drop the packets (usually called null routing). Null routes work great, but usually drop all packets to a particular destination IP, which means you need to move your service to another IP if you want it to stay online; that's hard to do if your IP needs to stay fixed for a meaningful time (TTL for glue records at TLDs is usually at least a day); and IP space is limited, so if your attackers are quick at moving attacks, you could run out of IPs to use. Some attacks are going above 1 Tbps though, so that's a lot of bandwidth if you need to accept and drop; and of course, the more bandwidth people get so they can weather attacks, the more bandwidth that can be used to attack others if it's not well secured.

Re: Cloudflare was down

#318
What was interesting and scary is that our monitoring system didn’t notify us. Our email was down because we use cloudflare for DNS and our monitoring provider’s SMS gateway was down. So we didn’t get sms messages.

Re: Cloudflare was down

#319

It really defies the original vision of the internet to have so many services depend on a single company. Almost every news site I was reading dropped off at once. I thought for a second that I lost internet in my own house.

I consider DNS and the way how top level domains are handled to be one of the weakest parts of our current Internet design. We REALLY need a truly decentralized, distributed DNS system that is not owned by private entities.

DNS worked just fine throughout this. You're barking up the wrong tree.

Re: Cloudflare was down

#320
post #153

Earlier quoted context omitted.

I've never even heard of the site before. Nor have I searched for "attorney" any time recently.

We operate that site and are using Cloudflare to prevent DDOS attacks. Probably some sort of hash collision...

Crazy stuff!
Post reply on HN