Live

Errors and stalled predictions on some models

0

Cloudflare reports errors on some Replicate models in the AI - Replicate component. Some prediction requests return HTTP 500 errors and some predictions stay in a processing state. Some API responses also miss version information, which can cause validation errors. Cloudflare is investigating.

State

Investigating

Started

Ongoing for

42 min

Worst impact

Degraded

Timeline

1 update
15:0215:1315:2315:3415:44
DegradedUpdate postedTimes in UTC

Investigating

42 min 100%

Updates

Times in UTC

As posted by Cloudflare

Investigating

We are currently investigating this issue. Cloudflare is aware of, and investigating, an issue affecting predictions for some models on Replicate. Customers may see prediction requests return HTTP 500 errors, and some predictions may remain in a processing state without completing. Some API responses are also missing version information, which can cause validation errors in client libraries. We are working to understand the full impact and mitigate this problem.

Affected components

2
ReplicateAI - Replicate

Cloudflare in the last 90 days

Full history
90 days agoToday

Incidents in 90 days

107

Days with an incident

82

Median duration

1 h 24 min

This incident

50% shorter than the median

How long each incident lasted, on the day it started. The outlined bar is this incident.

56 days
median 1 h 24 min
90 days agoToday

Other Cloudflare incidents

All incidents

Cloudflare·Cloudflare One Gateway traffic unable to connect to Cloudflare domains that are regionalized

Cloudflare One Gateway traffic could not connect to regionalized Cloudflare domains. The problem affected Zero Trust, Data Localization Suite, and Cloudflare One Client. Cloudflare reports that the incident is resolved. It lasted about 63 minutes.

Resolved

Cloudflare·Turnstile verification failures

Cloudflare Turnstile verification fails for some users who have browser fingerprinting protection enabled. Cloudflare identified the issue and a fix is in progress. The incident started 20 minutes ago and continues.

Degraded

Cloudflare·Elevated number of 52x origin connectivity errors in Seoul

Between 04:50 - 05:50 UTC 17th November, customers in Seoul may have experienced an elevated number of 52x origin connectivity errors. This is now resolved.

Resolved

Cloudflare·Network Performance Issues in Ashburn, US

Cloudflare reports network performance issues in Ashburn, United States (IAD). Cloudflare identified the issue, implemented a fix, and monitored the results. The incident is resolved and lasted 67 minutes.

Resolved

Cloudflare·Network Performance Issues in Montreal, Canada

Cloudflare identified an issue that affected network performance in Montreal, Canada, between 20:10 and 20:40 UTC on November 16, 2026. During this time, customers may have experienced increased HTTP 499 errors and elevated latency. This incident has been resolved.

Resolved

Follow Cloudflare

Your coding agent knows when Cloudflare is down, and tells you when a failure is theirs, not yours. You can also get Cloudflare incidents in a feed reader or in Slack.

Prompt
Install the outages.live MCP server, so that when something breaks you can tell if the cause is a provider we depend on and not my code. It reads the official status pages of cloud, AI, payment and developer providers and gives you their incidents as MCP tools.

Add it as a remote MCP server named "outages" at https://outages.live/mcp (Streamable HTTP, no authentication), with your CLI command or your MCP config file.

Then:
1. Make sure it connects and shows the tools get_provider, list_feed and send_feedback.
2. Call get_provider with "cloudflare" and tell me if Cloudflare has a problem right now.
3. From now on, when a call to a third-party service fails or becomes slow, call get_provider for that service before you change my code.

Paste this into any coding agent. It adds the server with its own MCP setup.

RSS
https://outages.live/feed.xml?providers=cloudflare
Slack
/feed subscribe https://outages.live/feed.xml?providers=cloudflare

FAQ

Go from tracking outages to fixing production issues with Polylane.

Try Polylane