Cloudflare favicon

Cloudflare outage on 2026-10-27

cloudflarestatus.com
October 2026 27 Major incident

Cloudflare logged eight separate incidents on 27 October 2026, starting at 08:04 UTC and running through the evening. The longest, a two hour 46 minute stretch of ERR_TIMED_OUT and ERR_EMPTY_RESPONSE errors for traffic passing through external vendor firewalls, was on Cloudflare's side, not yours.

Started

12:28 UTC

Duration

Lasted 2h 46m

Source

IsDown

Next time

Cloudflare is degraded. Is your product working?

An incident on their side does not always mean an outage on yours, and the only way to know is to already be watching your own endpoints. Set that up in under a minute and you will be notified the next time your own site is affected. And there will be a next time.

50 monitors for your own websites. No card.

Affected components: Cloudflare Sites and Services, Workers AI, Stream, Stockholm, Sweden - (ARN), CDN Cache Purge, Europe, Workers Builds, Pages

What happened?

Cloudflare's status page recorded eight incidents across the day, touching Pages, Workers Builds, the CDN cache purge system, Stream, Workers AI, Access, and network error logging, plus a Stockholm (ARN) location entry. The day opened at 08:04 UTC with Pages and Workers Builds intermittently failing because GitHub repository cloning was broken, fixed in 14 minutes. A second Workers Builds problem followed at 11:38, this time builds failing to deploy with Containers, lasting one hour 19 minutes. The longest incident started at 12:28: customers whose traffic routed through an external vendor firewall saw ERR_TIMED_OUT on standard HTTPS sites and ERR_EMPTY_RESPONSE on non-standard ports like 8443. Cloudflare said it was investigating root cause while working a fix, and reported the fix in place at 14:52, two hours 46 minutes after it started. A cache purge delay affecting the EEUR region ran from 15:10 to 16:17, with Cloudflare telling customers to use Purge Everything as a workaround. Stream had playback and upload issues from 16:10, AI Security for Apps had detection failures from 16:31, Access API errors ran from 21:13 to 21:58, and the day closed with Network Error Log processing down from roughly 19:30 UTC, discovered and announced at 22:21, running 37 minutes once addressed. Every incident here closed with a monitoring update rather than a final resolved note, so each fix is reported by Cloudflare but not confirmed closed in writing. No user reports were filed during this window on UptimeRobot's side. Cloudflare's own history shows 201 incidents in the past 90 days with a median duration of 1 hour 55 minutes, so a day with eight entries and one incident over two and a half hours sits within that provider's normal pattern, not outside it.

Learning

UptimeRobot's probes against Cloudflare's own control plane, DNS, and edge data components held 100% uptime over 24 hours and 7 days, and 99.996% over 30 days, through this entire day. That is not a contradiction. The incidents above were scoped failures, firewall interaction errors, build pipeline faults, a regional cache queue backing up, none of which necessarily register as a failed check against Cloudflare's core infrastructure from outside. A customer's product can be returning ERR_TIMED_OUT to its own visitors while the edge network that UptimeRobot checks responds normally, because the fault sits in a narrower path than the one being probed. This is the general shape of a Cloudflare incident: something specific breaks for a subset of traffic or a specific feature, Cloudflare posts it, and broad external monitoring often won't show it because it isn't broad external failure. If you only watch your own endpoints, a fault like the firewall interaction issue or the cache purge delay looks like your problem until you go looking elsewhere. Your monitoring covers your half, UptimeRobot now watches the provider's half too, so next time you know which side broke without guessing.

4.7
stars out of 5
284+ reviews on

Start monitoring in 30 seconds.

There's nothing to install. No credit card required. 50 monitors for free.