We have successfully mitigated the issue affecting the us-east1-a region. The facility power disruption has been addressed, and the impacted Infiniband networking fabric and associated systems have returned to normal operation. All affected services are now stable, and full functionality has been restored. Our teams will continue to monitor the region closely to ensure continued stability.
We appreciate your patience during this incident and apologize again for any disruption it may have caused.
Posted Aug 21, 2025 - 19:59 UTC
Update
We are continuing to investigate and mitigate the service degradation affecting our us-east1-a region, following a facility power disruption at our data center. Our teams remain in close coordination with the data center provider as they work to fully restore services. Recovery of critical systems remains our top priority.
We sincerely apologize for the ongoing impact and appreciate your continued patience as we work to resolve the issue.
Posted Aug 21, 2025 - 02:32 UTC
Investigating
We're investigating a service degradation in our us-east1-a region, triggered by a facility power disruption at our data center.
The primary impact is to the Infiniband networking fabric, which may cause intermittent errors or failures for multi-node, distributed workloads. Some customers may also experience individual virtual machines becoming unavailable. Our teams are working to identify all affected resources.
Our engineering teams are actively working to stabilize the affected systems and mitigate the risk of further disruption. We are coordinating with our data center provider to support their remediation efforts and restore full service resiliency as quickly as possible. We apologize for any impact this is causing.
Posted Aug 20, 2025 - 17:09 UTC
This incident affected: GPU Virtual Machines (us-east1) and Networking (Infiniband Networks).