Cloudflare outage history

From 1:15 PM UTC to 3:15 PM UTC, Cloudflare customers in France and Indonesia faced degraded functionality as Durable Objects and downstream services reported errors, with Jakarta datacenter requests experiencing disconnects. Separately, users in India encountered network performance issues. The impact was regional rather than global, affecting apps relying on durable objects and routing in these areas.

Cloudflare's updates documented ongoing investigation, identification of the issue, and fixes that were implemented with monitoring. By 3:15 PM UTC, affected network components had recovered to up, and Durable Objects issues in the FR/ID regions were stabilizing.

Impact began at 11:40 AM UTC with network performance issues in India, causing slower responses for Cloudflare sites and services. The incident expanded to France and Indonesia, where Durable Objects and downstream services reported increased errors, and connections to the Jakarta data center occasionally resulted in disconnections. Visitors in these regions likely saw degraded performance and intermittent failures as the incident evolved.

Cloudflare experienced a brief network performance issue in the Singapore region, impacting connectivity for users in that area. The incident began at 5:45 AM UTC as engineers started investigating the cause and mapping mitigations.

A fix was implemented by 5:55 AM UTC and Cloudflare continued to monitor the results. By 6:00 AM UTC, the status had returned to normal, with the impact contained toSingapore and no wider disruption.

Cloudflare reported elevated error rates in the Chicago ORD region starting at 9:20 PM UTC, affecting users whose traffic traversed Cloudflare's network through ORD. Sites and services relying on Cloudflare in that region experienced errors and slower responses during the incident. Cloudflare indicated they were investigating the elevated error rates in Chicago (ORD) while the ORD network showed degraded performance.

Cloudflare reports a brief disruption to the Workers AI GLM 5.2 component within Cloudflare Sites and Services, resulting in degraded Workers AI functionality. The incident began at 4:20 PM UTC and was resolved by 5:30 PM UTC after about an hour, as a fix was implemented and monitoring began.

In the Hong Kong region, Cloudflare Durable Objects and some related products (RealtimeKit and Realtime SFU) experienced an elevated error rate that affected users in the area. The issue began at 7:00 AM UTC, with Cloudflare noting an investigation and potential impact to multiple customers in the region. A fix was implemented by 7:10 AM UTC, and services began to recover as monitoring continued. By 7:25 AM UTC the Durable Objects component had returned to normal operation and status was updated to up, indicating the disruption was resolved.

The impact was localized to the Hong Kong region and affected components included Cloudflare Sites and Services / Durable Objects, with some related products such as RealtimeKit and Realtime SFU likely experiencing degraded functionality during the incident.

Cloudflare briefly experienced degraded performance as requests to CDNJS (cdnjs.cloudflare.com) encountered elevated errors. The incident began at 4:20 PM UTC with CDNJS-related errors under investigation and was resolved by 4:40 PM UTC after a brief investigation. This disruption affected access to JavaScript libraries hosted on cdnjs.cloudflare.com, with users potentially experiencing failed or slower resource loads during that window.

Impact began at 8:00 PM UTC when Cloudflare listed the Lists service as degraded. Changes to URL Lists were delayed, and edits could return errors, while existing Lists continued to function normally. The issue affected the Cloudflare Sites and Services / Lists component and impacted users managing URL lists.

A fix was implemented at 10:20 PM UTC and the service returned to normal by 10:45 PM UTC, with ongoing monitoring to confirm stability.

Cloudflare experienced network performance issues beginning at 7:35 PM UTC in Phoenix, with traffic from Los Angeles affected shortly after 7:40 PM UTC. The impact included degraded user connectivity across Cloudflare's regional paths. Later, around 8:00 PM UTC, the Lists service degraded due to delays in URL Lists changes, with existing Lists continuing to function. A fix for the Lists component was implemented by 10:20 PM UTC and monitoring began, with the Lists service returning to normal by 10:45 PM UTC.

Users attempting to load the Cloudflare Dashboard encountered errors starting around 5:30 PM UTC. Cloudflare noted that the issue did not affect the Cloudflare API, the CDN edge caching, or other security features at the edge; the impact was limited to the Dashboard experience. A fix was identified around 6:55 PM UTC and monitoring continued. The Dashboard returned to normal by 7:30 PM UTC, with the incident lasting roughly two hours for affected users.