On this page
The Cloudflare outage began at 11.20am UK time on Tuesday 18 November 2025. Most traffic was flowing again by 2.30pm and everything was back to normal by 5.06pm. An internal database change made a configuration file double in size, which crashed the software that handles traffic. X and ChatGPT were among the sites affected. Cloudflare said it was not a cyber attack.
- The outage began at 11:20 UTC on 18 November 2025.
- Core traffic recovered by 14:30 UTC and all systems by 17:06 UTC.
- An oversized internal file crashed Cloudflare's traffic software.
- Cloudflare said there was no cyber attack. It was not a power cut.
What happened at Cloudflare
Late on the morning of Tuesday 18 November 2025, websites across the internet began showing error pages. The Register reported that X, formerly Twitter, was knocked offline or misbehaving, and that the fault-tracking site Downdetector was itself showing an error at one point. The Register's own site was affected.
The sites had one thing in common. They all use Cloudflare, a company that sits between websites and their visitors to speed pages up and block unwanted traffic. When Cloudflare's network stopped passing traffic, the sites behind it could not be reached. It came less than a month after the AWS outage of 20 October 2025, which disrupted Lloyds, HMRC and Snapchat for much of a day.
When it started and how long it lasted
Cloudflare's account says its network began failing to deliver traffic at 11:20 UTC, which is UK time in November. For a while the network swung between working and failing, which made the cause harder to spot.
A corrected file was sent out at 14:30 UTC and most services recovered then, a little over three hours after the start. Cloudflare says all systems were working normally by 17:06 UTC, just under six hours in total.
Who and what was affected
Cloudflare listed its core website delivery and security services, its login check tool Turnstile, its dashboard and its Access sign-in service among those affected. Visitors to customer websites saw server error pages.
Cloudflare does not publish a list of affected customers. The Register described chunks of the web going offline worldwide. Any UK business, council or news site that uses Cloudflare could have been affected for those hours.
The cause, as stated by Cloudflare
Cloudflare traced the outage to a change in permissions on one of its databases, made at 11:05 UTC. The change caused the database to write duplicate entries into a file used by its Bot Management system, which decides whether a visitor is a person or an automated program.
The file roughly doubled in size and was copied to every machine on Cloudflare's network. The software reading it had a fixed limit on the file's size. When the larger file arrived, the software failed, and the websites it served returned errors. The pattern echoes the CrowdStrike outage of July 2024, when a faulty settings update sent to security software crashed an estimated 8.5 million Windows devices.
Cloudflare at first suspected a very large attack, partly because its separate status page happened to go down at the same moment. That turned out to be a coincidence. The company said no cyber attack or malicious activity of any kind was involved. A year later the Microsoft 365 outage of November 2026 was likewise traced to a hidden software defect and not to an attack, though it lasted just under 24 hours.
What customers were told
Cloudflare acknowledged the problem on its status page at 11:48 UTC and said at 13:09 UTC that it had found the cause. It apologised to its customers and to internet users in general, The Register reported.
The company said it would harden the way its own configuration files are checked, add more switches to turn features off quickly, and review how its core software fails. A separate and smaller incident followed in the Cloudflare outage of February 2026, when a bug in a clean-up task withdrew the internet routes for about 1,100 blocks of customer addresses.
What this does and does not tell us about power cuts
This was an internal software fault. It was not a power cut, it did not cause one, and Cloudflare ruled out an attack. No official body has said a cyber attack on the UK electricity grid is expected.
The lesson for a household is modest. A website being down, even a lot of websites at once, is usually an ordinary fault that is fixed within hours. If you are unsure what is going on, a broadcast source is a good cross-check, and our guide to getting news in a blackout explains why the Be Ready campaign suggests a battery or wind-up radio.
For money, the simple backup is a small amount of notes and coins, and our page on keeping cash for an emergency explains how much is sensible.
Common questions
Was the November 2025 Cloudflare outage a cyber attack?
No. Cloudflare first suspected one but concluded the cause was an internal database change that produced an oversized file.
Which sites went down in the Cloudflare outage?
The Register reported that X and Downdetector were among those affected, along with many other sites that use Cloudflare.
How long did the Cloudflare outage last?
Most traffic recovered a little over three hours after the 11:20 UTC start. Cloudflare says all systems were normal by 17:06 UTC.
- Cloudflare: Cloudflare outage on November 18, 2025
- The Register: Cloudflare coughs, half the internet catches a cold
Reviewed against these sources on 18 November 2026. If official advice has changed since, follow the official advice.
General information only. This page is not professional, safety, medical, legal or financial advice, and it is not official guidance. No national blackout has been forecast or announced, and nothing here suggests one is expected. We are independent and have no connection to the government or any energy company. Always check the official sources linked on this page. Full disclaimer
Keep reading
Amazon Web Services outage, October 2025: what happened and why
The Amazon Web Services outage of 20 October 2025 disrupted Lloyds, HMRC, Snapchat and many more. AWS blamed a DNS fault in its DynamoDB database service.
Read: Amazon Web Services outage, October 2025:CrowdStrike outage, July 2024: what happened and why
The CrowdStrike outage of 19 July 2024 crashed about 8.5 million Windows computers worldwide, hitting GP surgeries and airlines. What happened and why.
Read: CrowdStrike outage, July 2024: what happenedMicrosoft 365 outage, November 2026: what happened and why
The Microsoft 365 outage of 12 and 13 November 2026 cut off Outlook email worldwide for almost a day. What failed, why, and what it means for UK households.
Read: Microsoft 365 outage, November 2026: whatCloudflare outage, February 2026: what happened and why
The Cloudflare outage of 20 February 2026 took some customers' websites offline for about six hours. Cloudflare blamed a bug in an automated clean-up task.
Read: Cloudflare outage, February 2026: what