Cloudflare said it is investigating an outage affecting its global network services after users reported “internal server error” messages when attempting to reach affected websites and online platforms. The company first acknowledged the issues in a status post, noting availability problems with its support portal.
Cloudflare’s Global Network spans servers and data centers in more than 330 cities across over 120 countries, providing content delivery, security and performance services. The company reported a global network edge capacity of 449 Tbps and said it connects to over 13,000 networks, including major ISPs, cloud providers and enterprises.
At 11:48 UTC Cloudflare added a further incident post for the Global Network and said it was investigating an issue impacting multiple customers, including widespread 500 errors and failures of the Cloudflare dashboard and API. Cloudflare said it was working to understand the impact and mitigate the problem.
In tests conducted during the outage, Cloudflare nodes across Europe were reported down in cities including Bucharest, Zurich, Warsaw, Oslo, Amsterdam, Berlin, Frankfurt, Vienna, Stockholm and Hamburg. Outage monitoring service Downdetector has received tens of thousands of reports, with users experiencing issues with server connections, websites and hosting services have also reported issues.
Cloudflare has previously handled other large incidents, including a June outage that affected Zero Trust WARP connectivity and Access authentication across regions and that also impacted Google Cloud infrastructure, and an October outage tied to a major DNS failure that disrupted millions of websites on a major cloud provider. The company did not provide a technical cause for the current outage in its initial posts.
Later updates indicated signs of recovery: an update at 07:29 EST said services were showing recovery but that customers might still see higher-than-normal error rates, and a subsequent update at 08:47 EST said changes had allowed Cloudflare Access and WARP to recover and that WARP access had been re-enabled in London while work continued to restore application services.