Cloudflare centralizes its outage tracking in the Radar Outage Center

Cloudflare has long documented major Internet disruptions—from government-directed shutdowns in Iran and Sudan to infrastructure failures like the October 2021 Facebook outage—through blog posts and, more recently, near-real-time updates on its @CloudflareRadar Twitter account. With the release of Radar 2.0, the company is consolidating that reporting into a single, searchable resource: the Cloudflare Radar Outage Center (CROC).

The CROC is designed as an archive of observed outage events, organized by location, type, date, and other attributes. Cloudflare intends it to become a primary reference for civil society groups, journalists, and affected parties seeking confirmation or additional context on reported outages.

What the CROC records

Each entry in the CROC summarizes a single observed outage event with the following fields:

  • Location: The geographic area where the outage occurred.
  • ASN: The autonomous system that experienced the connectivity disruption.
  • Type: The breadth of the outage—national, sub-national, or network-level.
  • Scope: The specific state, city, or network provider affected.
  • Cause: Likely reason based on publicly available information, such as government-directed shutdowns, weather events, cable cuts, power failures, or filtering/blocking.
  • Start time: When the outage began.
  • End time: When the outage resolved.

Each listing links to supporting material—a tweet, blog post, or a custom Radar graph—for users who want more detail on a particular event.

Finding and using outage data

Outage information is surfaced in a few ways. Radar pages, including the main landing page, display a card with the most recent observed outage and a link to the CROC. The CROC itself is also accessible from Radar’s left-side navigation. Inside the tool, a global map shows locations with outages during the selected time period, with tooltips indicating the number of events, and a table below lists each outage’s details.

The CROC is built on the same API that Cloudflare launched alongside Radar 2.0, meaning the underlying data is programmatically accessible. API documentation is available in the Cloudflare API documentation for researchers and data journalists who want to integrate outage records into their own systems. For those who prefer offline analysis, a "Download CSV" link under the table exports the entire outage list.

Roadmap: from Internet-scale to cloud-scale monitoring

The initial CROC release is a foundation rather than a finished product. Cloudflare plans to automate more of its outage detection so that events are published sooner through both the API and the web interface. The team is also investigating the use of synthetic monitoring, combined with network-level performance data, to identify outages affecting major consumer and business applications.

One motivating problem is the well-known sluggishness of cloud platform providers' own status pages. It is common for users to see widespread reports of a service failure on social media while the provider's status page still shows "All systems normal." Cloudflare says it has spent recent months testing whether its own monitoring can reliably detect cloud provider outages, with the goal of extending the CROC to include those providers. If successful, the company says the result would give the Internet a real-time status page it currently lacks, by tracking services whose official status pages often lag behind reality.