Nebius Outage History

84 incidents recorded over the last 365 days, 77 of them resolved. Durations are measured from when PulsAPI first saw the problem to when it cleared, which is usually longer than the vendor's own figure.

Live Nebius status100.00% uptime history

Every Recorded Nebius Incident

Nebius incidents, newest first, with date, duration and status.
DateIncidentDurationStatus
Nov 2, 2026Public API gateway partially unavailable in US-NORTH1 and US-CENTRAL1This incident has been resolved.3h 5mResolved
Oct 31, 2026Token Factory: some dedicated endpoints partially unavailableThis incident has been resolved.7h 17mResolved
Oct 28, 2026Compute Cloud: disk images and snapshots partial unavailabilityThe incident has been resolved6h 54mResolved
Oct 27, 2026Scheduled Network Maintenance Between eu-west1 and eu-north1Maintenance is complete. All services are fully operational.6h 0mScheduled
Oct 24, 2026Token Factory 503 errors in eu-west2The issue has been resolved.1h 2mResolved
Oct 18, 2026Partial IAM and console unavailabilityThe console and IAM were unavailable for approximately 10 minutes. The incident has been mitigated, and service availability has been restored.0 minResolved
Oct 16, 2026CLI login issuesThis incident has been resolved.10 minResolved
Oct 15, 2026InfiniBand connectivity issues for GPU clusters in eu-west2This incident has been resolved.10h 38mResolved
Oct 14, 2026Connectivity issues between eu-south1 and other Nebius regionsThis incident has been resolved.9h 9mResolved
Oct 14, 2026InfiniBand connectivity issues for GPU clusters in eu-west2This incident has been resolved.58 minResolved
Oct 8, 2026Metrics unavailable in Management Console for Monitoring in eu-west1The incident has been resolved.28 minResolved
Oct 7, 2026Problems with power for several dozens of nodesThis incident has been resolved.14h 11mResolved
Oct 3, 2026InfiniBand connectivity issues for GPU clusters in eu-west2This incident has been resolved.4h 13mResolved
Oct 2, 2026Issues with preempted Managed Kubernetes nodes in all regionsThis incident has been resolved.72h 22mResolved
Oct 2, 2026Compute VM creation and Object storage are partial unavailableThis incident has been resolved.32 minResolved
Sep 23, 2026Network issues in eu-west2This incident has been resolved.3h 11mResolved
Sep 22, 2026Object storage unavailable in eu-west1Incident is resolved.0 minResolved
Sep 17, 2026Compute services are partially unavailable for customersThe root cause was found and this incident has been resolved.43 minResolved
Sep 17, 2026Resuming IAM accounts is unavailableThis incident has been resolved.58 minResolved
Sep 16, 2026[eu-west1] Metrics for resources in this region are partially unavailableThis incident has been resolved.1h 28mResolved
Sep 16, 2026[us-central1] Object Storage is partially unavailableThis incident has been resolved.4h 56mResolved
Sep 14, 2026Managed Service for SkyPilot: cluster creation and update requests failing in eu-north1This incident has been resolved.24h 33mResolved
Sep 9, 2026Unstable compute VM createThis incident has been resolved.24 minResolved
Sep 1, 2026Connectivity issues for newly created worker nodesThis incident has been resolved.15 minResolved
Aug 19, 2026Region us-central1 partial degradation[https://nebius.com/blog/posts/incident-post-mortem-analysis-us-central1-service-disruption-on-august-19](https://nebius.com/blog/posts/incident-post-mortem-analysis-us-central1-service-disruption-on-august-19)21h 32mResolved
Aug 17, 2026Issues logging to the Management ConsoleSome newly registered users may have experienced issues logging in to the management console between 13:00 UTC on 17 August 2026 and 14:00 UTC on 18 August 2026.0 minResolved
Aug 15, 2026Partial InfiniBand unavailabilityThis incident has been resolved.3h 17mResolved
Aug 12, 2026External VPC connectivity issuesThis incident has been resolved.25 minResolved
Aug 5, 2026Creation of new VMs is brokenThis incident has been resolved.1h 18mResolved
Jul 27, 2026Network issuesThis incident has been resolved.1h 33mResolved
Jul 27, 2026Scheduled firmware upgrade in uk-south1 affecting B300-based instancesThe scheduled maintenance affecting B300 GPU workloads in the uk-south1-a IB fabric has been completed successfully. Thank you for your patience and understanding. If you have any questions or need assistance, please…72h 0mScheduled
Jul 22, 2026[eu-north1] Network timeouts affecting multiple servicesThe eu-north1 region experienced network issues that may have affected the operation of several services. The issue has been resolved, and all services are operating normally.0 minResolved
Jul 14, 2026InfiniBand switch failure affecting GPU workloads in the eu-north1-c fabric in the eu-north1 regionThis incident has been resolved.22 minResolved
Jul 3, 2026Managed SkyPilot OIDC Login is unavailableManaged SkyPilot OIDC Login was unavailable since 2 of July. The problem has been identified and resolved.0 minResolved
Jul 1, 2026Monitoring, Object Storage and Token Factory services are partially unavailableThis incident has been resolved.1h 43mResolved
Jun 30, 2026New virtual machines are not being createdThis incident has been resolved.4 minResolved
Jun 24, 2026Object Storage service partially unavailable in US-CENTRAL1 regionThe Object Storage service in the US-CENTRAL1 region has been fully restored and is operating normally. The root cause has been identified, and corrective measures are in place.1h 36mResolved
Jun 16, 2026One InfiniBand switch failureThis incident has been resolved.32 minResolved
Jun 15, 2026Network Connectivity IssuesConnectivity issues are gone now. Intermittent failure is being investigated in the background with the uplink provider, while the Nebius Netowrking team continues to monitor the link closely.21 minResolved
Jun 10, 2026Compute instances startup issuesThis incident has been resolved.2h 32mResolved
Jun 9, 2026Some public IPs unreachable in eu-north1Some of public IPs became unreachable.-175 minResolved
Jun 4, 2026Network Connectivity IssuesThis incident has been resolved.1h 28mResolved
Jun 2, 2026Unable to allocate new public IPv4 addresses in eu-north1This incident has been resolved.24h 12mResolved
May 29, 2026Power issuesSeveral servers were rebooted.-58 minResolved
May 19, 2026Unable to perform operations on a small subset of virtual machinesThis incident has been resolved.1h 54mResolved
May 13, 2026InfiniBand performance degradationThis incident has been resolved.2h 19mResolved
May 7, 2026Nebius console is unavailableIssue resolved 21:10.24 minResolved
May 6, 2026Managed Kubernetes operations can take longer than expectedThis incident has been resolved.19h 6mResolved
Apr 29, 2026Partial unavailability of external models in Token FactoryThe external provider has restored operations. All models are now available.16h 28mResolved
Apr 24, 2026Partial degradation of Dedicated Endpoints in Token FactoryThis incident has been resolved.43 minResolved
Apr 20, 2026Metrics for VMs are unavailableSince 8:30 back to normal.25 minResolved
Apr 20, 2026TokenFactory DataLab unavailabilityThis incident has been resolved.25h 48mResolved
Apr 15, 2026TokenFactory — Some Serverless Models Partially Unavailable in us-central1 RegionThis incident has been resolved.1h 36mResolved
Apr 10, 2026Degraded logs, metrics and audit trails availabilityDuring the night we experienced issues with storage service. While writing for data was not affected (just delayed), but reading logs, accessing metrics and audit trails in console was not possible while the incident…0 minResolved
Apr 7, 2026Observability data partially unavailable for Dedicated Endpoints in TokenFactoryThis incident has been resolved.1h 14mResolved
Apr 1, 2026Slower project and tenant creationThis incident has been resolved.59 minResolved
Mar 26, 2026Scheduled power infrastructure maintenance in us-central1 affecting H200-based instances.The scheduled maintenance has been completed.19h 30mScheduled
Mar 26, 2026Managed k8s is partially unavailable for customersA fix has been implemented and we are monitoring the results36 minResolved
Mar 24, 2026Storage Metrics Ingestion FailureWe had a problem with metrics ingestion which resulted in storage metrics not being ingested during the following intervals: - eu-west: from 14:50 CET to 19:05 CET - me-west1: from 17:00 CET to 19:05 CET - eu-west1…0 minResolved
Mar 23, 2026Observability data partially unavailable in TokenFactoryThis incident has been resolved.1h 26mResolved
Mar 10, 2026Power issues during planned maintenanceThis incident has been resolved and all services and clusters should be operational now.9h 2mResolved
Mar 10, 2026Urgent Maintenance: Power Supply Works for GPU H200 NodesThe scheduled maintenance has been completed. All power supply works on the H200 nodes have been finished, and all nodes in the cluster have been fully restored.15h 0mScheduled
Mar 5, 2026Some VMs scheduling can take more than 45 minutesThis incident has been resolved.41 minResolved
Mar 2, 2026Power issues in us-central1This incident has been resolved.1h 17mResolved
Feb 26, 2026Power issues in eu-north1This incident has been resolved.10h 15mResolved
Feb 20, 2026Network issues in us-central1This incident has been resolved.55 minResolved
Feb 18, 2026Batch inference unavailability in Token FactoryThis incident has been resolved.21h 34mResolved
Feb 18, 2026Continuation of Scheduled InfiniBand MaintenanceMaintenance is over and the team is monitoring the situation closely.4h 0mScheduled
Feb 18, 2026Partial unavailability of Managed Kubernetes LoadBalancersThis incident has been resolved.3h 5mResolved
Feb 18, 2026Scheduled InfiniBand MaintenanceThe scheduled maintenance has been completed.4h 0mScheduled
Feb 16, 2026One Infiniband switch failureThis incident has been resolved.17 minResolved
Feb 10, 2026Partial loss of Infiniband connectivityThe connectivity is fully restored.25 minResolved
Feb 9, 2026Partial unavailability of Console functionalityThis incident has been resolved.31 minResolved
Feb 9, 2026Planned Maintenance: IB Switches Reboot in fabric-7 in eu-north1The scheduled maintenance has been completed.2h 0mScheduled
Feb 6, 2026Partial degradation of dedicated endpoints in Token FactoryThis incident has been resolved.1h 35mResolved
Feb 6, 2026Partial degradation of dedicated endpoints in Token FactoryThe incident has been resolved.3h 17mResolved
Feb 6, 2026TokenFactory Files API / Batch API and new LoRA deployments are unavailableThe incident has been resolved.40 minResolved
Feb 2, 2026Compute services are partially unavailable for customersAt 10:20 GMT all failed workloads were recovered.30 minResolved
Jan 28, 2026Token Factory control plane is experiencing partial downtime. Inference requests are not impacted.This incident has been resolved.1h 54mResolved
Jan 22, 2026Partial monitoring unavailabilityThis incident has been resolved.1h 45mResolved
Jan 6, 2026Consumption data is not updating for all users.This incident has been resolved.5h 14mResolved
Jan 5, 2026A subset of metrics related to storage and observability systems was not collectedThis incident has been resolved.14 minResolved
Dec 30, 2025Partial unavailability of the InfiniBand fabric fabric-7This incident has been resolved.45 minResolved
Dec 24, 2025[UK-SOUTH1] Partial unavailability of the InfiniBand factoryOur team detected an issue with an InfiniBand switch in the fabric, which caused a partial network interruption for a subset of nodes/workloads. The switch was successfully rebooted and connectivity has been restored…0 minResolved

Hear about the next one while it is open

Everything above is a record written after the fact. We can email you when the next Nebius incident opens instead, at the point where knowing still changes what you do.

Free, no password, no card. Unsubscribe in one click.

How to Read This Nebius Incident Log

Each row is an incident PulsAPI observed, not a summary written afterwards. The duration is wall-clock time between the first failing check and the first clean one, so it includes the window before Nebius acknowledged anything. Vendor post-mortems typically measure from acknowledgement, which is why their numbers are usually shorter.

Incidents still open have no duration yet and are listed as ongoing rather than being given a running total. The archive covers the last 365 days; anything older has aged out of the window rather than never having happened.

The longest single Nebius outage in this window ran 3d, against a mean recovery of 5h 57m. If you depend on Nebius in a customer-facing path, the longest figure is the one to design around. The mean is what happens on a normal bad day; the maximum is what happens on the worst one.