That's a lot of website that are now down... I find it impressive how terrible bigger services that are supposedly "redundant" are at reliability because of all the added complexity.
It happens... It's happened with S3, other AWS, Azure, GCP, and a number of other systems over the years. Seems to happen a couple times a year between them all.
That said, on self-hosting I've had outages that were self-inflicted more often than with cloud pipelines and CI/CD flows to production... in the end, it's probably a wash or favors cloud all the same. Even if it's more generally felt from cloud providers all at once.
Only your site is down, for the first time this year? You must suck at reliability!
Your site is down for the third time this year because of a Cloudflare or AWS outage, together with hundreds of others? Force majeure, can’t be helped.
If you are able to build a competing product that can offer what cloudflare does at the scale that it does I promise there is billions of dollars in it for you.
I'm really unsure what you are trying to say. Is there someone doing what cloudflare is doing now? I do not think so. Until that day, this will be how it is. It's very silly to complain about "centralization" when there is really not an alternative to what cloudflare provides.
Why is this level of centralization accepted? Many European governments have also started using Cloudflare after suffering DDoS attacks following the 2022 Russian invasion of Ukraine and European sanctions on Russia.
No, because it's complicated to run, and difficult to the point of impossible for a normal small amount of servers to withstand a DDOS. Why do we get our electricity from the power company, even though there are occasionally outages? We could just all get generators and battery-backed solar, but that would be expensive and hard to maintain. The scale of AWS, Cloudflare, Akamai, etc. make them less bad at it that the average company that needs a website. It's the same reason we don't all have our own mailservers.
It would be nice to have a choice, but most of their competition is either priced completely out of the reach of most people, or PAYG with unbounded spending so you can get hit with an enormous bill out of nowhere, which you don't have to worry about with Cloudflare. AWS or Azure or Google could make their pricing structures as approachable as CFs if they wanted to, and peel off many of CFs small-to-medium scale customers, but evidently none of them want to even try. They only care about scoring whales.
Because being down when nobody else is will be seen as your lack of preparedness, and being down together with everybody else doesn’t look half bad in comparison.
ugh, this again. Honestly you should be able to answer your own question at this point. At least say something original like how a small-time developer can get a planet wide delivery network, DDoS mitigation, and regional traffic routing without relying on something like Cloudflare.
I noticed my website has been timing out randomly, and checking Apache server-status page I don't see the requests that are in flight.
At first when I noticed timeouts and slow response, I assumed it was due to my connection or my server. However, after checking on apache, it seems some requests aren't going through at all.
I checked the CF status page at the time, there wasn't this notice, but now there is. I hope this gets resolved soon.
I had to change my DNS from Cloudflare to Google last night somewhere around midnight Eastern because everything but telegram stopped working, presuming that Telegram must use an IP or something for its endpoints.
darn, i wish this script had each location resolve it's own dns, that way could gain a glimpse of global external dns cloudflare querying issues if you mtr a name that uses CF...
