- Google Cloud suspended Railway’s account following suspicious activity
- The hours-long outage affected all Railway workloads across all clouds
- Railway is taking ownership of the oversight due to a technical dependency
Platform-as-a-Service (PaaS) company Railway has accused Google Cloud of abruptly suspending its account without prior warning, leading to an hours-long outage.
The company, which counts over three million users who host around 10 million services, APIs and databases, identified an issue at around 10:20pm UTC on May 19 2026, which was only fully rectified around eight hours later at 6:14am.
While the company determined the cause and sought to reinstate services, customers were subject to errors like ‘no healthy upstream’, ‘unconditional drop overload’, login failures and an inability to access the dashboard.
Technical dependency ultimately caused Railway’s outage
Because the network control plane API was affected (which is hosted on Google Cloud), all of Railway workloads across all clouds were affected, returning 503 and 404 errors. Existing workloads stayed up and running for around 15 minutes before caches started to expire.
The Register reports Railway spends an eight-figure sum every year with Google Cloud (potentially upwards of $1m every month), even after having moved parts of its infrastructure into colocation services following 2024 and 2025 issues.
Railway claims it took Google Cloud nearly an hour to engage after the incident happened. “We are livid and still trying to get all the details,” Solutions Engineer Angelo Saraceno said. It’s worth noting that the account got unsuspended by 10:29pm, just nine minutes after issues began.
The company has since published a detailed blog post uncovering everything it knows about the incident, patched together with information from Google Cloud.
The report confirms that the outage was part of a broader, automated sweep by Google that impacted other Google Cloud Platform accounts without prior warning. TechRadar Pro has learned that Google Cloud identified a surge in abusive activity, particularly cryptocurrency mining, across a large number of accounts, and that it had previously warned affected users of suspicious activity and potential lock-outs accordingly.
Exclusive dependency is a bad idea
Railway has taken full ownership of its oversight and has announced immediate changes, including removing the network plane API’s sole dependency on GCP.
“If any of the interconnects go out, there is always a path between the clouds,” Support Engineer Chandrika Khanduri and Agent Experience GM Cody De Arkland explained.
“We have invested in resiliency as a result of prior incidents which have assisted us in dealing with the impact,” the company added, hinting at previous issues. Though it promises to have learned from previous mistakes, it was ultimately Railway’s responsibility to eliminate its GCP-only dependency, which could have prevented widespread impact this time around.
“Your customers don’t care whether the failure was Google or Railway; they see your product. Your uptime is our responsibility, and we’ll keep delivering on it,” the company concluded.
Follow TechRadar on Google News and add us as a preferred source to get our expert news, reviews, and opinion in your feeds.
https://cdn.mos.cms.futurecdn.net/f2noH8daNe2WNYuaTKbjxT-1267-80.jpg
Source link
