Cloudflare favicon

Cloudflare outage on 2026-11-16

cloudflarestatus.com
November 2026 16 Minor incident

On 16 November 2026, Cloudflare had a 1h 49m outage affecting prediction creation for some models on Replicate, which runs on Cloudflare's infrastructure. If your product calls those models, requests failed with "Failed to create prediction" during the window, and the fault sat on Cloudflare's side.

Started

14:30 UTC

Duration

Lasted 1h 49m

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.

What happened?

Cloudflare's status page opened an incident at 14:30 UTC titled "Errors creating predictions for some models". The first update said Cloudflare was aware of and investigating an issue affecting prediction creation for some models on Replicate, with customers seeing requests fail with the error "Failed to create prediction". Cloudflare said it was working to understand the full impact and would post more updates shortly. At 15:30 UTC, Cloudflare said the issue had been identified and a fix was being implemented. At 15:52 UTC, Cloudflare said a fix had been implemented and it was monitoring results. No further update followed after that, and the incident is recorded as resolved after 1h 49m. Cloudflare filed three updates in total, and this page shows all three. No components are listed as affected in Cloudflare's own breakdown for this incident. No user reports were filed during the window on this page's record. Cloudflare has not published a post-mortem explaining what caused the prediction creation failures, so the root cause is not on the record. Over the past 90 days, Cloudflare has logged 209 incidents with a median duration of 1 hour 55 minutes, so this outage ran close to Cloudflare's own typical length. Across all time, Cloudflare's status page has recorded 3,510 incidents.

Learning

UptimeRobot's own probes against api-control, dns-1111 and edge-data, run from EU and NA regions at 60 second intervals, show 100% uptime over the last 24 hours, 7 days and 30 days, and this incident was not registered against those checks. That is not a contradiction. The failure here sat specifically in prediction creation for some models on Replicate, a narrower function than the general availability that edge and DNS probes are built to catch, so a clean external read and a real failure for affected requests can both be true at the same time. This is the shape of failure worth recognising: the parts of a platform you can ping from outside stay up while a specific feature path inside it breaks for whoever depends on it. Watching your own endpoints would only tell you that your side is responding, not that a call downstream to Replicate's prediction creation was returning errors. Over 30 days, UptimeRobot's own stats show 77 incidents tracked against this provider with a median duration of 1h 37m, giving you a baseline to judge future outages against. 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.