Web infrastructure company Cloudflare says all its services are back online after a major outage on Tuesday disrupted traffic to some of the world’s most visited websites.
The incident affected users globally for several hours before engineers applied a fix and began monitoring recovery.
According to News24, the outage began at around 13:30 SAST and hit businesses and consumers worldwide, with X, ChatGPT, Canva and Spotify among the services that were intermittently unavailable.
Cloudflare, which is responsible for handling about a fifth of global web traffic, confirmed it was experiencing an internal service degradation and warned that some services might be impacted while it investigated.
Cloudflare’s status page shows that engineers first flagged the problem at 11:48 UTC (13:48 SAST), reporting an internal service degradation and warning of intermittent impact on services as they focused on restoring normal operations.
Subsequent updates noted elevated error rates and latency across the global network as traffic routing and application services were affected.
The company reported that it had identified the issue at 13:09 UTC (15:09 SAST) and started deploying a fix.
By 14:42 UTC (16:42 SAST) Cloudflare said the fix had been implemented and that it believed the incident was resolved, although monitoring would continue to ensure that all services had fully recovered.
At 17:44 UTC (19:44 SAST) the company confirmed that services were operating normally and that it was no longer observing elevated errors or latency, while a deeper investigation into the disruption was under way.
During the incident, some customers struggled to log into or use the Cloudflare dashboard, and there were temporary restrictions on WARP access in London as part of the remediation steps.
Cloudflare also warned that bot scores used by its security tools could be intermittently affected while systems recovered across the network.
News24 reports that financial markets reacted to the outage, with Reuters noting a 4.1% fall in Cloudflare’s share price in premarket trading.
Downdetector recorded more than 5,600 reports of problems with X in the United States just before 07:00 EST as users struggled to load feeds or post content.
The outage came shortly after other large scale incidents at major cloud providers. In October, separate outages at Amazon Web Services and Microsoft Azure disrupted services for millions of users globally.
Cloudflare has continued with planned maintenance work on several of its regional data centres following the recovery.
Status notices show scheduled maintenance in Hong Kong and Kuala Lumpur from 18:00 to 22:00 UTC (20:00 – 00:00 SAST), Sydney from 15:00 to 19:00 UTC (17:00 – 21:00 SAST), and Atlanta between 07:00 UTC (09:00 SAST) on 18 November and 22:00 UTC (00:00 SAST) on 19 November.
During these windows, the company warned that traffic might be rerouted, that users in the affected regions could experience slightly higher latency, and that some private and cloud network interfaces might temporarily fail over to other locations.
The company has indicated that a full post incident report will be published once its internal investigation into the root cause of Tuesday’s disruption is complete.
