Cloudflare Outage History

Cloudflare faced a widespread disruption beginning on Feb 16, 2026 10:20 AM UTC, with Newark (EWR) datacenter latency and intermittent errors affecting a subset of HTTP traffic. Over the next days the issue expanded to increased HTTP 5xx responses for R2 access from APAC, elevated errors in Mumbai, and intermittent WARP connectivity problems for iOS users. The disruption affected multiple Cloudflare components, including R2, Bot Management, SSL certificate provisioning, WARP, and Durable Objects, leading to timeouts and degraded edge services for many customers. The outage spanned from Feb 16, 2026 10:20 AM UTC to Feb 20, 2026 11:05 PM UTC with ongoing mitigation efforts documented on the status page.

Cloudflare provided a steady stream of updates throughout the incident, documenting investigations, identified fixes, and monitoring results. Regional activity covered Mumbai, APAC regions, and other datacenters such as Sofia and Bangalore as Cloudflare worked to restore full service. The broad scope and multi-day duration indicate a major disruption across a wide range of Cloudflare services during this period.

Cloudflare conducted scheduled maintenance affecting data center regions in Sofia, Belgrade, and Düsseldorf. The maintenance began at Feb 15, 2026 11:00 PM UTC and was expected to run through Feb 16, 2026 8:00 AM UTC. During this window, users in these regions may have experienced intermittent performance or partial service disruption as services were updated and traffic was routed through maintenance tasks.

The incident involved four official status updates across the regions, with Sofia first, followed by Belgrade and Düsseldorf, and a final update when the status was restored to up. The maintenance concluded by Feb 16, 2026 8:00 AM UTC. A small number of user reports were logged during the window, indicating limited impact.

Cloudflare reported a brief disruption impacting Stream Live playback, causing buffering and stalling during live broadcasts while on-demand videos and Media Transformations were unaffected. The incident began at Feb 15, 2026 6:20 PM UTC and was resolved by Feb 15, 2026 6:50 PM UTC after Cloudflare identified a fix and monitored results. The Stream component under Cloudflare Sites and Services experienced a temporary degradation with three status updates issued during the incident. User reports were limited, and there was no observed impact to other Cloudflare services.

Cloudflare experienced a multi-service degradation beginning with scheduled maintenance in IAD (Ashburn) on Feb 13, 2026 5:00 AM UTC. The disruption affected AI Gateway, Dashboard, and Load Balancing and Monitoring, causing elevated 500 errors for Unified Billing credits, 429 errors on the Dashboard, and analytics delays in Load Balancing. A separate maintenance entry for HKG (Hong Kong) appeared later in the timeline. A fix was implemented around 8:15 PM UTC and services began recovering; Dashboard errors were noted to be resolving around 8:52 PM UTC. The incident concluded with a return to normal by Feb 14, 2026 12:35 AM UTC after roughly 19 hours and 35 minutes of degraded performance. The Cloudflare status page posted eight updates during the incident.

Cloudflare reported scheduled maintenance in the Portland area began at 2:00 PM UTC. The incident progressed with updates indicating Sippy 404 errors on R2 objects that exist upstream, followed by reports of high error rates on R2 FedRAMP buckets. The service maintained a degraded level through the event, issuing four official status updates as teams investigated and mitigated the impact, with the issue resolving by 1:05 AM UTC.

The incident began on Feb 10, 2026 2:00 PM UTC with a Portland scheduled maintenance notice and quickly evolved into broader degraded performance affecting multiple Cloudflare products. By Feb 10, 2026 6:35 PM UTC, users reported elevated error rates across Workers AI, Stream, Images, and Firewall for AI, with AI Search also impacted. Later updates indicated high error rates for R2 FedRAMP Buckets, signaling a wider impact on data services. Regional maintenance events were noted in Riga around Feb 10, 2026 10:45 PM UTC and other locations as the incident unfolded. Over the next two days Cloudflare issued 11 official status updates describing ongoing investigations, maintenance windows, and issues across multiple components, including Zone Versioning analytics, the Dashboard SSO connector, and Turnstile. Some issues began to resolve gradually, for example Turnstile widget validation failures were addressed and Turnstile came back up by Feb 12, 2026 1:50 PM UTC, while others remained degraded or undergoing maintenance during the period. Regional maintenance activity occurred in Las Vegas around Feb 11, 2026 9:00 AM UTC and Zagreb around Feb 12, 2026 2:00 AM UTC, illustrating the geographic spread of the event. The incident spanned from Feb 10, 2026 2:00 PM UTC to Feb 12, 2026 1:50 PM UTC, with partial recoveries in several components by the end of the period.

Cloudflare's WARP service experienced issues that could show an untrusted SSL certificate warning for users visiting msn.com during a sequence of regional maintenance events. Portland maintenance began at Feb 9, 2026 2:00 PM UTC, followed by Dublin maintenance at Feb 10, 2026 2:30 AM UTC and Rio de Janeiro maintenance at Feb 10, 2026 7:00 AM UTC. The incidents affected Cloudflare Sites and Services / WARP and Cloudflare Sites and Services / Workers KV, with SSL warnings reported by users as the main symptom. The event timeline spans from Feb 9, 2026 2:00 PM UTC to Feb 10, 2026 12:00 PM UTC and included multiple status updates (six in total).

Cloudflare conducted scheduled maintenance in the EZE (Buenos Aires) region. The maintenance window began at 7:00 AM UTC and the status page indicated the event as in progress with a scheduled maintenance notice, affecting the Buenos Aires edge region.

The maintenance concluded at 1:05 PM UTC, lasting approximately 6 hours and 5 minutes. Cloudflare issued 2 official status updates during the incident. Geographic impact was limited to the EZE region.

Cloudflare reported a regional incident in the Dallas, Texas area with a surge in HTTP 500 errors affecting traffic routed through Cloudflare's network. Users accessing services via Dallas endpoints experienced degraded performance and failed requests during the incident. The outage began at 2:15 AM UTC and lasted about 2 hours and 30 minutes, with service restored by 4:45 AM UTC.

Cloudflare conducted scheduled maintenance in the NRT (Tokyo) region on 2026-02-06. The maintenance caused a brief degradation and a warning status while work was in progress. The maintenance began around 4:30 PM UTC and the issue was resolved with service restored by 9:05 PM UTC, lasting about 4 hours and 30 minutes. The service posted two status updates during the incident to inform users. The impact appeared localized to the Tokyo region during the maintenance window.