Cloudflare outage history

Users in the ENAM region experienced degraded performance beginning at 8:10 PM UTC as Cloudflare flagged an uptick in errors across Durable Objects and downstream services. This affected several Cloudflare Sites and Services components, notably AI Search, Artifacts, Containers, D1, Durable Objects, Workers Assets, and Workflows, causing errors and slower responses for resources built on these features. Cloudflare indicated the issue was under investigation and began mitigation efforts as the situation evolved.

A fix was implemented by 8:55 PM UTC and Cloudflare started monitoring results to stabilize the region. By 11:10 PM UTC the affected components had returned to normal, and Cloudflare resumed typical operations with ongoing monitoring.

The incident began at 3:00 PM UTC when Cloudflare flagged MCP Portal tool invocation logging instability, causing some logs to be dropped and affecting logpush and the dashboard. Portal connectivity and initial authentication logs were not impacted. The issue was confined to internal accounts and did not cause customer-facing outages. The disruption lasted roughly forty minutes and was resolved by 3:40 PM UTC. A follow-up at 3:30 PM UTC confirmed the impact remained isolated to internal accounts with no customer-facing effects, with the primary affected component being Cloudflare Sites and Services / MCP Portal.

Users experienced degraded performance due to increased latency across Cloudflare Sites and Services / CDN/Cache affecting a subset of requests. The issue began at 8:40 PM UTC and persisted for most of the period, with a small volume of user reports noting slower responses.

At 2:45 PM UTC, Cloudflare said a fix had been implemented and they were monitoring the results. The service returned to normal shortly after, with status updated to up at 2:55 PM UTC.

Cloudflare's SSL Certificate Provisioning component entered maintenance at 12:05 PM UTC and returned to normal at 1:05 PM UTC. This brief maintenance window affected the certificate provisioning workflow rather than causing a full outage; users may have experienced delays in SSL-related provisioning tasks during that hour while the component was in maintenance.

Cloudflare reported a regional disruption beginning in Chicago, where HTTP 5xx errors increased and affected CDN/Cache services. The issue started at 5:50 AM UTC and cloud services users may have observed failed or slow requests while the provider investigated. By 6:55 AM UTC, a separate note indicated that Durable Objects (DO) and D1 services were degraded in Hong Kong and Singapore, with a fix implemented and monitoring underway. In the following hours, additional components began returning to normal, and by 8:30 AM UTC multiple key services—Durable Objects, D1, and R2—were back to up, signaling a broader restoration across Cloudflare’s platform.

Impact began around 3:55 AM UTC, with Hong Kong and Singapore experiencing degraded operations tied to Durable Objects, D1, and R2 in Cloudflare Sites and Services. At about 5:50 AM UTC, Chicago saw an uptick in HTTP 5xx errors affecting the CDN/Cache edge, indicating a broader but partial disruption across Cloudflare’s services. A fix for the HK/SG degradation was implemented around 6:55 AM UTC and monitoring followed. By 8:30 AM UTC, the affected components had returned to normal status and the Chicago edge issues appeared resolved as well.

In the Dubai region, Cloudflare experienced elevated errors for R2 requests, affecting Cloudflare Sites and Services / R2. The issue began at 2:35 AM UTC and was resolved by 3:00 AM UTC, with R2 operations returning to normal about 25 minutes later. The impact appeared localized to Dubai (DXB), where users relying on R2 may have encountered failed or slower R2 operations during the window.

The incident was brief, with Cloudflare restoring normal service after the transient degradation. No broader global impact was indicated beyond the Dubai region.

An incident lasting about 50 minutes began at 8:45 PM UTC when Cloudflare's Zero Trust posture checks started failing for some customer configurations, causing delays when accessing resources protected by WARP posture rules. The issue appeared to be tied to a timeout with a customer-configured provider, with the provider timing out when posture check requests were sent. The service returned to normal by 9:35 PM UTC as posture checks resumed, and users experienced degraded performance rather than a full outage during the event.

Users in the Asia Pacific region experienced degraded performance and intermittent connectivity to sites and services that rely on Cloudflare, starting at 5:40 PM UTC. Cloudflare acknowledged an ongoing investigation into APAC network performance and began working to mitigate the problem.

At 5:50 PM UTC a fix had been implemented and monitoring of results began. The issue was resolved by 7:45 PM UTC and service returned to normal, indicating restoration of normal operations within the APAC region.

Users in the Cloudflare Hong Kong region experienced elevated errors and higher latency affecting Durable Objects, R2, and related RealtimeKit features, leading to degraded performance for applications relying on these components. Cloudflare acknowledged the issue and began investigating at 3:05 PM UTC, with a fix implemented by 3:15 PM UTC and monitoring ongoing. All affected components returned to normal by 3:25 PM UTC; the issue appears localized to the Hong Kong region with no confirmed wider impact.