Railway Platform Hit by 8-Hour Google Cloud Account Suspension
Original: Incident Report: Railway Blocked by Google Cloud (Resolved)
Why This Matters
Highlights risks of single cloud provider dependency in platform architecture
Railway experienced platform-wide outage on May 19-20, 2026 when Google Cloud incorrectly suspended their production account for 8 hours. The incident affected all Railway workloads across regions, causing 503 and 404 errors for users attempting to access the platform.
Railway reported a major service disruption lasting approximately 8 hours from May 19 22:20 UTC to May 20 06:14 UTC when Google Cloud incorrectly placed their production account in suspended status. The suspension immediately took down Railway's API, control plane, and databases hosted on GCP. While workloads on Railway Metal and AWS remained operational initially, the outage cascaded as Railway's edge proxies rely on GCP-hosted control plane APIs for routing tables. As cached network routes expired, all Railway workloads became unreachable with 404 errors. Recovery began at 22:29 UTC when GCP access was restored, but persistent disks and networking took hours to fully recover. The incident also triggered GitHub rate-limiting on Railway's OAuth integrations and reset terms-of-service acceptance records. Railway takes responsibility for architectural decisions that allowed single provider failure to cascade platform-wide.