Many large websites went down Tuesday, including Amazon, Twitch, BBC, NYT, Spotify, Reddit, and FT, due to an issue at the Fastly CDN, which has now been fixed

TechCrunch 2021-06-09

Context & Ripple Effects

The outage exposed how a single delivery-network failure can interrupt consumer, media, and commerce services at once. Fastly later attributed the incident to a bug triggered by a customer configuration, turning a broad availability failure into a question of platform safeguards.

It fits a related run of shared-infrastructure disruptions: an earlier AWS outage affecting multiple online services and a later Akamai Edge DNS disruption likewise spread one provider problem across otherwise unrelated brands.

First-order effects

Amazon, Twitch, BBC, NYT, Spotify, Reddit, and FT regain service after Fastly fixes the issue, while Fastly must address the failure path that let one configuration trigger widespread disruption.

Second-order effects

Fastly customers have a concrete reason to review how much of their public availability depends on one CDN and how configuration changes are isolated before deployment.
Akamai and AWS face sharper scrutiny from customers evaluating whether their own concentrated infrastructure dependencies have comparable outage paths.

Third-order effects

Repeated failures across Fastly, AWS, and Akamai point to internet reliability becoming a supply-chain risk: services may compete independently yet share a small set of infrastructure providers.
If this pattern persists, resilience architecture and change-control practices become a more important basis for selecting cloud, DNS, and CDN vendors than feature breadth alone.

The trend: Major online services are becoming more visibly dependent on a small number of shared infrastructure layers, making provider outages cross-industry events.

Discussion

Vox Rebecca Heilweil on x
A company you've probably never heard of caused half the internet to go dark
@fastly @fastly on x
We identified a service configuration that triggered disruptions across our POPs globally and have disabled that configuration. Our global network is coming back online. Continued status is available at https://status.fastly.com/...
@tim Tim Bradshaw on x
One innocent customer's action took out 85% of the Fastly network yesterday. A configuration change triggered a previously unknown, month-old software bug. It was (mostly) fixed in 49 minutes. https://twitter.com/...
@ap @ap on x
Fastly said: “The issue has been identified and a fix has been applied. Customers may experience increased origin load as global services return.” A number of sites that were hit early appeared to be coming back online. https://apne.ws/jv3mi7a
@thestalwart Joe Weisenthal on x
$FSLY +3% as everyone realized that if a $6 billion company could take down the internet then it should be worth more than $6 billion.
@googledocs @googledocs on x
Glad we could help out! https://twitter.com/...
@kevincollier Kevin Collier on x
The Verge is publishing through a live Google doc because fastly went down this morning, then this happened. Our whole online information infrastructure is so much more tenuous and ridiculous than we usually consider. https://twitter.com/...
@halsparks Hal Sparks on x
Imagine if a pinched nerve could make your whole body collapse and stop working... https://twitter.com/... https://twitter.com/...
@ariehkovler Arieh Kovler on x
When there's a HUGE outage, it's usually a config error. https://twitter.com/...
@arrington @arrington on x
This right here is why I'm bullish on Cloudflare @eastdakota @Cloudflare https://techcrunch.com/...
@govuk @govuk on x
We are aware of the issues with https://gov.uk/ which means that users may not be able to access the site. This is a wider issue affecting a number of other non-government sites. We are investigating this as a matter of urgency.
@charlescwcooke Charles C. W. Cooke on x
There's really no point in having a distributed Internet if every major player is going to put their infrastructure behind the same proxy. https://techcrunch.com/...
@jennelizabethj @jennelizabethj on x
Let's not centralize the internet so much, shall we? A blip this time, catastrophic next time. #internetdown https://www.theguardian.com/ ...
@fabiochiusi Fabio Chiusi on x
“Despite speculation on social media that the outage was the result of a malicious attack, leading to the hashtag #cyberattack trending on Twitter, there is no evidence pointing to foul play” https://www.theguardian.com/ ... https://twitter.com/...
@kendilaniannbc Ken Dilanian on x
This just in: The internet is important to our daily lives. If this can happen by accident, imagine what an adversary could do. https://twitter.com/...
@digiphile Alex Howard on x
Good morning! Content delivery networks (CDNs) closely resemble critical infrastructure in the 21st century for digital governments. @fastly told the @guardian a “configuration error” was at fault: https://www.theguardian.com/ ... “there is no evidence pointing to foul play”-@ale…
@tomwarren Tom Warren on x
look ma, my mistake made some news https://metro.co.uk/...
@kjhealy Kieran Healy on x
The whole Internet loves Decentralized Packet Switching Network, a lovely topology with no central points of failure! *50 years later* We regret to inform you
@journodave David Wyllie on x
excited to start my job as chief marketing officer of Fastly, with my pitched campaign of ‘everyone should know who we are and what we do’ time to sit down and take a big sip of cof
@jessicaelgot Jessica Elgot on x
Luckily WhatsApp is working, so Westminster and the government are still functioning
@ap @ap on x
Numerous websites were unavailable on Tuesday after an apparent widespread at cloud service company Fastly. Dozens of high-traffic websites including the New York Times, CNN, Twitch and the U.K. government's home page, could not be reached. https://apne.ws/VlyKzjo
@pwnallthethings @pwnallthethings on x
We did it everyone, we finally ran out of internet
@verge @verge on x
And we're back! Normal service on the internet has resumed, with the outage traced back to CDN provider Fastly. We'll see you all again in Google Docs for the next disaster. https://www.theverge.com/...
@rustybrick Barry Schwartz on x
Hey @johnmu - does GoogleBot go into special relax mode and take a bit of a vacation during these outages? https://twitter.com/...
@rustybrick Barry Schwartz on x
CNN looking good https://twitter.com/... https://twitter.com/...
@iblametom Thomas Brewster on x
Thank the lord TechCrunch has survived https://twitter.com/...
@brianstelter Brian Stelter on x
Regarding this morning's widespread website outages: “Typically an outage of this nature occurs when a vital internet infrastructure service faces an issue. An apparent glitch at Fastly, a popular CDN provider, is thought to be the reason...” https://techcrunch.com/...
@ruskin147 Rory Cellan-Jones on x
Masses of sites down, including https://gov.uk/ and many news sites
@campuscodi Catalin Cimpanu on x
Major internet outage developing after Fastly went down earlier https://twitter.com/...

Chronicles