Railway: How a GCP account suspension took Railway down for 8 hours
On May 19, 2026, Railway experienced a platform-wide service disruption lasting approximately eight hours. The SEV-1 outage began when Google Cloud incorrectly placed Railway’s production account into a suspended status. This action instantly disabled Railway’s GCP-hosted infrastructure, taking down the dashboard, API, databases, compute infrastructure, and critical network components. While Railway operates a multi-cloud architecture spanning GCP, AWS, and Railway Metal, the outage quickly cascaded globally. A hidden architectural dependency—a GCP-hosted network control plane—prevented edge proxies from refreshing routing tables. Once cached routes expired, customer services across all cloud providers became unreachable. This incident serves as a stark reminder that multi-cloud infrastructure does not guarantee resilience if a critical control-plane dependency relies on a single provider.
Railway