OVHcloud Outage History

OVHcloud experienced a network incident in the GRA11 region from 21:40 to 22:55 UTC on March 8, 2026, causing temporary network instability for services in that region. The incident was caused by an unexpected underlying infrastructure malfunction and lasted approximately 75 minutes before being resolved.

OVHcloud experienced a major incident affecting some compute instances in the GRA1 region, making them unreachable due to an unexpected underlying infrastructure malfunction. Customers were temporarily unable to access and use their affected instances in the region. The incident was identified and teams were mobilized to restore service, with the issue lasting approximately 4 minutes.

OVHcloud's Object Storage S3 Standard service in the SBG region experienced a major outage from 07:30 to 12:30 UTC on March 6, 2026, caused by an electrical equipment issue. Customers were unable to access, download, or upload S3 objects due to 500/503 errors during the 5-hour incident. The issue was identified and resolved by OVHcloud's technical teams, restoring full service functionality.

OVHcloud experienced a minor incident affecting their Managed Kubernetes Service across multiple regions (AP-SOUTH-MUM, EU-WEST-PAR, EU-SOUTH-MIL, and RBX) starting at 09:00 UTC on March 5, 2026. The incident impacted the Containers & Orchestration service offering for customers using Kubernetes clusters in these regions. The issue was resolved within 7 minutes after investigation teams addressed the underlying problem.

OVHcloud experienced an API incident in the SBG5 region affecting their Compute - Instance service from March 4, 2026 at 22:15 UTC to March 5, 2026 at 00:00 UTC (lasting 24 minutes). The incident prevented customers from creating, starting, or managing instances and creating or attaching new volumes, though existing running instances remained unaffected. The issue was resolved after teams identified and fixed an unexpected underlying infrastructure malfunction.

OVHcloud's AI & Machine Learning service experienced a major incident affecting the LLama3.3-70B model from March 4-11, 2026, causing significant latency issues and intermittent unavailability for customers. The incident was initially caused by unusually high load on the model, which led to broader software issues that required extensive remediation efforts. The service was temporarily restored on March 6th but experienced additional instability before being fully resolved on March 11th after 166.7 hours of total disruption.

OVHcloud experienced a major incident affecting compute instances in the Gravelines (GRA3) region from March 3, 2026 16:33 UTC to 18:45 UTC, caused by an unexpected infrastructure malfunction. Some instances became unreachable, preventing customers from accessing and using their instances in the affected region. The incident was resolved after approximately 2 hours and 12 minutes, with most instances restored partway through the outage.

OVHcloud's AI & Machine Learning service experienced a major incident affecting the Mistral model endpoints, causing slowness and unavailability for customers using this AI model. The incident was caused by a software issue and lasted 52.6 hours from February 26, 2026 at 12:00 UTC until March 4, 2026 at 16:36 UTC. Service was fully restored on March 3rd with continued monitoring to ensure stability.

OVHcloud's Object Storage service in the EU-WEST-PAR region experienced a major outage from 09:00 to 15:30 UTC on February 28, 2026, causing 503 errors and preventing customers from uploading or downloading objects across all three availability zones (A, B, and C). The incident was caused by an unexpected underlying infrastructure malfunction and was resolved after approximately 6.5 hours of investigation and remediation efforts.

OVHcloud experienced a network equipment failure that affected their API Network for Public Cloud services, including Data & Analytics and Managed Kubernetes Service in the GRA9 region. The incident lasted 43 minutes from 14:09 to 14:52 UTC on February 27, 2026, and was resolved after teams addressed the underlying network equipment issue.