Cloudflare outage, 21 June 2022

Cloudflare BGP policy change takes 19 data centers offline (21 June 2022). The disruption lasted about 1 hour 15 minutes, according to Cloudflare’s own timeline.

What happened. A network configuration change to BGP prefix advertisement policies, part of a resilience project, was rolled out to the spines of 19 Multi-Colo PoP data centers. A re-ordering of policy terms withdrew a critical subset of prefixes, taking those locations offline; they handled about 50% of total requests.

What changes. Cloudflare will add MCP-specific test/deploy procedures, redesign the policy statement to prevent mis-ordering, and automate staggered rollouts with automated ‘commit-confirm’ rollback.

Questions

What caused the Cloudflare outage in June 2022?

A network configuration change to BGP prefix advertisement policies, part of a resilience project, was rolled out to the spines of 19 Multi-Colo PoP data centers.

Which services were affected?

Cloudflare network (CDN/proxy and services in affected locations).

Read how shared dependencies spread a failure in why one cloud outage can break unrelated apps, or browse the outage archive.