Health ( Europe Brief News): A major Google Cloud outage disrupted AI services like Vertex AI and Dialogflow CX, impacting healthcare systems in the U.S., Asia, and Europe.
Multiple Google Cloud services, such as Google Vertex AI Online Prediction, Dialogflow CX, Agent Assist, and Contact Center AI, were unavailable due to a problem impacting users in the U.S., Southeast Asia, and Europe for three hours and 53 minutes.
“To our customers who were impacted during this disruption, we sincerely apologize. This is not the level of quality and reliability we strive to offer you, and we are taking immediate steps to improve the platform’s performance and availability,”
Google said in a statement.
The technology giant reported that the primary reason for the outage was that on June 8, users initiated a new category of requests, leading to sporadic crashes (termed “segmentation fault”) on certain servers managing AI responses.
If a server failed, it would automatically reboot, and incoming requests would be sent to other functional servers. At first, the requests weren’t an issue since there were sufficient healthy servers to manage the demand, Google stated.
Nonetheless, as the volume of these crash-inducing requests rose, an excessive number of servers failed simultaneously, resulting in insufficient available servers to meet the demand. Consequently, users began to experience outages.
On June 10, Google engineers were notified of the problems, and at that point, no observable impact on customers was detected. The engineers discovered the issue and started implementing a solution.
Today, many services utilizing Vertex Prediction started to encounter user-facing problems related to the same underlying issue. Google engineers found a link to the previous issue and sped up the implementation of the solution, fixing the problem.
Google announced plans to improve production monitoring to identify early indications of server failures and to reinforce its validation process for feature modifications and updates prior to their deployment in production, in order to avert the problem from happening again.
Hospitals, digital health startups, research institutions, and pharmaceutical firms use Vertex AI for diagnostic assistance, customized treatment suggestions based on patient data, risk assessment, and operational aid.
Dialogflow CX and Agent Assist are increasingly utilized in healthcare as tools for clinical support and to help with administrative tasks. Contact Centre AI is widely utilized in healthcare for scheduling patients, triage, billing assistance, and virtual entry services.
How might this outage impact AI-based healthcare services relying on Google Cloud?
Healthcare AI applications—like diagnostics, patient data management, telemedicine, and decision support—rely on constant cloud accessibility.
The disruption resulted in decreased performance and increased error rates in the impacted areas, causing interruptions or delays in these essential services.
AI systems in healthcare typically need immediate access to extensive datasets saved on Google Cloud Storage or analyzed using AI platforms such as Vertex AI. The disruption impacted these services, possibly hindering data collection, examination, and AI model inference vital for patient treatment.
Disruptions in AI-powered tools may postpone diagnoses, treatment suggestions, or monitoring notifications, potentially impacting patient results. Healthcare providers depending on cloud-based AI might have needed to return to manual methods or experience reduced service quality during the outage.
