OVHcloud outage history

OVHcloud scheduled electrical maintenance on Rack GRA0401B04B, with the GRA Bare Metal Services placed in maintenance. The status page indicated no impact on overall service, and warned that if a disruption occurred it could take up to 30 minutes to restore. Interruptions could occur at various times during the window and would not affect all racks simultaneously. The maintenance window began at 5:35 AM UTC and focused on the specified rack. By 10:19 AM UTC, the maintenance had concluded and the rack returned to normal operation. No service issues were observed during this period.

OVHcloud’s US-East NYC Public Cloud services experienced a resource constraint that halted reboot, rescue, and spawning operations for US-EAST-LZ-NYC-VPS-1, leaving NYC workloads temporarily unavailable. The outage began at 10:39 PM UTC and persisted into the following day as the NYC public cloud component recovered by 2:29 PM UTC. In parallel, YYZ Bare Metal Services were placed into maintenance starting at 11:05 AM UTC, including an electrical maintenance window for Rack YYZ0101A01A with an expected downtime of about 15 minutes; the maintenance window was not expected to impact all racks simultaneously. The YYZ rack completed maintenance and returned to normal by 4:15 PM UTC.

During a rolling upgrade of OVHcloud's ESXi hosts to ESXi 8.0U3k, Bare Metal Services began a maintenance window at 10:34 AM UTC. The upgrade was designed to be transparent, with running virtual machines migrated between hosts via vMotion and hosts rebooted one by one until the cluster was updated. In the afternoon, a PCI host in US-EAST-VA-1 went down, causing some VPS instances to be unavailable or unresponsive at 1:09 PM UTC. The operation briefly returned to maintenance at 1:14 PM UTC as the upgrade continued, and services returned to normal by 4:39 PM UTC for Bare Metal Services. The impact was localized to the US-East region, with other regions remaining operational.

OVHcloud reported a brief outage of a VPS host in the US-EAST-VA-2 region, causing some VPS instances in that area to be unavailable or unresponsive. The disruption began at 3:04 AM UTC and was resolved by 3:44 AM UTC, lasting roughly 40 minutes. OVHcloud indicated that teams were actively working toward a resolution, with updates focused on the US-EAST-VA-2 / Public Cloud Services component. The incident appears localized to that region and did not indicate broader regional impact.

OVHcloud conducted a planned ESXi upgrade maintenance to version 8.0U3k for VIN Bare Metal Services and VIN Network Services. The maintenance started at 12:04 AM UTC and was designed to be transparent, with each ESXi host placed into Maintenance Mode and running virtual machines vMotioned to other hosts. The upgrade proceeded on a rolling basis, upgrading one host at a time to minimize disruption.

During the window, Network Services began returning to normal by 3:05 AM UTC and Bare Metal Services followed, returning to normal by 3:34 AM UTC. By the end of the maintenance window, all VIN components were back online and operating as expected. The maintenance was communicated as a routine upgrade that aimed to minimize user impact by keeping workloads running with live migrations.

An EU-wide issue caused Windows license delivery to fail with an Error status due to an internal server error, affecting Windows license provisioning across Bare Metal regions WAW, GRA, LIM, RBX, and ERI. The impact was felt by customers attempting to obtain Windows licenses in the EU region. The incident began at 7:00 PM UTC with the status page indicating an investigation, and was resolved by 7:15 PM UTC as the license delivery flow recovered across the affected regions.

OVHcloud faced a partial network outage in its Vint Hill data center (Rack V113C07) that left some servers inaccessible via IPMI. The issue began at 8:30 AM UTC and was resolved by 10:04 AM UTC, with support monitoring for stability after restoration. The interruption mainly affected VIN / Network Services in that data center, within a broader period of ongoing infrastructure work across OVHcloud’sVIN and related racks.

In the days that followed, OVHcloud conducted a series of maintenance activities and ESXi upgrades across multiple racks (including HIL, VIN, and BHS). These actions included rack maintenance that produced brief downtimes of about 15 minutes and a rolling ESXi upgrade to version 8.0U3k. The upgrade process began around 6:24 PM UTC and continued through mid-August, with key private cloud services in the HIL and VIN regions returning to normal by 5:34 PM UTC as systems were upgraded and stabilized.

Separately, a VPS service issue in US-WEST-LZ-PAO-VPS-1 emerged on 4:44 PM UTC, causing delays and intermittent unresponsiveness for VPS tasks in that region. The situation gradually improved by 6:04 PM UTC. User reports from Europe and other regions also indicated VPS-related disruptions during this window, reflecting regionally distributed impact tied to the ongoing maintenance and upgrades.

OVHcloud carried out electrical infrastructure maintenance for Rack BHS0814B02C, affecting the BHS / Bare Metal Services. The work triggered a brief outage during the maintenance window, with approximately 10 minutes of downtime expected for affected services. The maintenance began at 2:35 AM UTC and, after completion of the work, normal service resumed by 11:04 AM UTC.

OVHcloud carried out electrical infrastructure maintenance on Rack BHS0812B04C, with Racks in maintenance: BHS0812B04C. During the window, Bare Metal Services on that rack experienced approximately five minutes of downtime, distributed across the window and not affecting all racks simultaneously. The rack returned to normal operation at 3:24 PM UTC after the maintenance began at 12:29 PM UTC.

OVHcloud reported a storage cluster problem affecting VPS services in the US-EAST-VA-2 region, with VIN / Virtual Private Services identified as the impacted component. Users could experience unresponsive instances when rebooting or attempting actions that change a VPS's state, and creation of new disks in this region was intermittently affected. The disruption began at 2:29 AM UTC as the issue emerged and OVHcloud worked on a fix.