Cloudflare Outage History

The outage began on Oct 10, 2025 8:00 AM UTC with network performance issues in five China datacenters (PKX, KWE, CZX, DLC, CKG), likely causing slower performance and intermittent connectivity for services routed through those locations. The situation expanded over the day, and around Oct 10, 2025 1:20 PM UTC Cloudflare noted a scheduled maintenance window in the LAX region as part of ongoing infrastructure activity. The incident persisted for about 24 hours and 15 minutes, with multiple updates issued to guide affected users.

Cloudflare experienced a partial outage in the Cloudflare Dashboard service, with requests to the Dashboard potentially failing and errors shown to some users. This incident affected the Cloudflare Sites and Services / Dashboard component, causing degraded dashboard access and user experience. The outage began at 4:30 AM UTC and was resolved by 7:10 AM UTC, lasting about 2 hours and 40 minutes. Cloudflare issued 2 status updates during the incident as they investigated and restored the Dashboard service.

The Cloudflare Marseille region (MRS) underwent scheduled maintenance beginning around 12:00 AM UTC. The status page indicated the work was in progress, which led to intermittent connectivity or degraded performance for users in the Marseille area. The maintenance lasted approximately 4 hours and 5 minutes, with services returning to normal by 4:05 AM UTC.

Cloudflare published two official status updates during the incident. There was no Early Warning Signal detected for this event, and the impact was regional to the Marseille data center.

Cloudflare reported a partial outage affecting Cloudflare Workers cron triggers, with a subset of scheduled tasks not being invoked. The disruption was limited to cron-based functions and did not indicate a platform-wide outage. The incident began around 9:12 PM UTC, with the status page updating to warning at 9:15 PM UTC. The issue was resolved by 10:55 PM UTC, lasting roughly 1 hour and 40 minutes. Cloudflare issued two official status updates during the incident, and the affected component was Cloudflare Sites and Services / Workers.

From 4:35 PM UTC the outage began with Cloudflare WARP connectivity issues, causing degraded Internet experience for WARP and Cloudflare Zero Trust users. The disruption expanded to other Cloudflare services, with reports of dropped connections for Stream Live, Realtime, and TURN, and broader performance problems across real-time services and secure access. The incident touched a wide set of Cloudflare components and affected users around the globe.

Cloudflare issued a continuous stream of status updates as the incident evolved, moving from Investigating to Update/Identified and Monitoring while engineers worked on mitigations. The issue affected multiple components including Realtime TURN Service, RealtimeKit, Stream, and related services such as API, Dashboard, Workers, and Bot Management. The outage persisted for over a week and a half, with the environment gradually recovering, and was considered resolved by 9:05 AM UTC.

User reports included connectivity issues from locations such as Phoenix, Montreal, Roanoke, and Perth, reflecting a widespread geographic footprint.

Cloudflare reported a brief outage affecting Billing operations, including the Cloudflare Dashboard and related APIs. Users attempting billing tasks may have encountered errors or unexpected behavior, while the CDN edge and other security features remained unaffected. The outage began at 11:10 AM UTC and was resolved by 11:35 AM UTC, lasting approximately 25 minutes.

Cloudflare issued two status updates during the incident, describing the issue and the subsequent fix. The provider stated that a fix had been implemented and monitoring the results. No other Cloudflare services, including cached content delivery via the CDN or edge security features, were affected.

Cloudflare experienced degraded performance beginning with Salt Lake City maintenance at Sep 29, 2025 8:05 AM UTC, which caused Log Explorer data processing delays and HTTP 500 errors for a subset of Load Balancing requests. Users attempting to query recent logs saw slower results, and some requests failed, though no logs were reported lost. The impact centered on two areas: Cloudflare Logs (Log Explorer and Logs) and the Cloudflare Load Balancing and Monitoring service.

The incident lasted about 25 hours, from Sep 29, 2025 8:05 AM UTC to Sep 30, 2025 9:10 AM UTC, and the provider issued 6 status updates detailing progress. By Sep 30, 2025 2:05 AM UTC the Load Balancing component had recovered, with Log Explorer delays gradually improving and full recovery achieved by Sep 30, 2025 9:10 AM UTC. All affected components returned to normal operation.

Cloudflare reported issues affecting Cloudflare Pages API, Workers Builds, and the SSL API (including SSL for SaaS Provisioning). The outage began around 3:50 AM UTC and lasted about one hour, with increased failures seen in Pages deployments and Worker builds. Cloudflare noted that all traffic to Workers and Pages would continue to function while deployments encountered failures, and the incident was resolved by 4:50 AM UTC. During the event, Cloudflare issued three status updates communicating investigation progress and eventual recovery.

Cloudflare reported a degradation near San Jose (SJC) impacting multiple services, including R2, D1, Durable Objects, Stream, and Cache Purge. Users may have experienced errors or timeouts when accessing resources routed through Cloudflare near the SJC region. The outage began at Sep 28, 2025 5:50 PM UTC and was resolved by Sep 28, 2025 9:30 PM UTC, lasting roughly 3 hours and 40 minutes. The provider issued two status updates during the incident as they investigated and then restored affected services.

All listed components returned to normal by the end of the incident, with R2, CDN Cache Purge, D1, Durable Objects, and Stream back in service by Sep 28, 2025 9:30 PM UTC.

Cloudflare reported a brief degradation of Cloudflare R2 in Baghdad (BGW), with elevated error rates affecting services in that region. The incident began on 9:05 PM UTC and was resolved by 9:35 PM UTC, lasting about 30 minutes. Cloudflare stated they were investigating the cause and applying mitigations, with the R2 component returning to normal after the incident window.