“We are furious and still trying to get all the details”: Google Cloud accused of suspending a major customer without any warning, leading to a major outage

Google Cloud suspended Railway account following suspicious activity
The hours-long outage affected all rail workloads across all clouds.
Railway takes ownership of monitoring due to technical dependency

The Railway-as-a-Service (PaaS) company has accused Google Cloud of abruptly suspending its account without prior warning, leading to an hours-long outage.

The company, which has more than three million users hosting approximately 10 million services, APIs and databases, identified an issue around 10:20 p.m. UTC on May 19, 2026, which was not fully fixed until approximately eight hours later, at 6:14 a.m.

While the company determined the cause and sought to restore services, customers were subject to errors such as “no healthy upstream”, “unconditional delete overload”, connection failures and inability to access the dashboard.

Technical dependence ultimately caused the railway to break down

Because the Network Control Plane API was affected (which is hosted on Google Cloud), all rail workloads across all clouds were affected, returning 503 and 404 errors. Existing workloads remained operational for approximately 15 minutes before caches began to expire.

The register reports that Railway spends an eight-figure sum with Google Cloud each year (potentially more than $1 million each month), even after moving some of its infrastructure to colocation services following the 2024 and 2025 issues.

Railway says it took Google Cloud almost an hour to engage after the incident. “We are furious and still trying to get all the details,” said solutions engineer Angelo Saraceno. It’s worth noting that the account suspension was reinstated at 10:29 p.m., just nine minutes after the issues began.

The company has since published a detailed blog post revealing everything it knows about the incident, corrected with information from Google Cloud.

The report confirms that the outage was part of a broader automated sweep by Google that affected other Google Cloud Platform accounts without prior warning. TechRadar Pro has learned that Google Cloud has identified an increase in abusive activity, particularly cryptocurrency mining, across a large number of accounts, and has previously warned affected users of suspicious activity and potential blocks as a result.

Exclusive dependence is a bad idea

Railway has taken full ownership of its oversight and announced immediate changes, including removing the Network Plan API’s sole reliance on GCP.

“If one of the interconnects goes down, there is always a path between the clouds,” explained support engineer Chandrika Khanduri and general manager of agent experience Cody De Arkland.

“We have invested in resilience following previous incidents, which has helped us deal with their impact,” the company added, alluding to previous problems. While he promises to have learned from previous mistakes, it was ultimately Railway’s responsibility to eliminate its reliance solely on GCP, which could have avoided widespread impact this time around.

“Your customers don’t care if the outage is Google or Railway; they see your product. Your uptime is our responsibility, and we will continue to respond,” the company concluded.

Google logo on black background next to text