Is CoreWeave down?

We are investigating an issue that is causing errors when creating new Dedicated Inference deployments or editing existing ones. Deployments that are already running are not affected and continue to serve traffic normally.

We have identified the cause of the errors affecting Dedicated Inference deployment creates and edits, and a fix is being prepared. Running deployments remain unaffected and are serving traffic normally.

A fix has been applied and Dedicated Inference deployment creates and edits are succeeding again. We are monitoring to confirm full recovery. If you are still seeing errors, please reach out to our support team.

Hello, We are currently investigating an issue affecting CoreWeave Cloud authentication (OIDC) services. Customers may experience issues accessing CAIOS, authenticating through OIDC, or obtaining credentials for CAIOS workloads. Our team is actively working with the highest priority level to restore full functionality as quickly as possible. We apologise for any inconvenience and appreciate your patience while we work to resolve the issue. — CoreWeave Support

Hello, We are currently investigating an issue affecting CoreWeave Cloud authentication (OIDC) services. Customers may experience issues accessing CoreWeave AI Object Storage (CAIOS), authenticating through OIDC, or obtaining credentials for CAIOS workloads. Our team is actively working with the highest priority level to restore full functionality as quickly as possible. We apologise for any inconvenience and appreciate your patience while we work to resolve the issue. — CoreWeave Support

Hello, The issue has been mitigated. CoreWeave Cloud authentication (OIDC) access to CoreWeave AI Object Storage (CAIOS) has returned to baseline, and affected workloads should now operate normally. Our teams continue to monitor this to prevent any recurrence. Thank you for your patience during this window. Thank You, CoreWeave Support Team.

Hello, We are investigating degraded performance on the backend GPU interconnect fabric in US-CENTRAL-08A. Multi-node GPU workloads may see reduced bandwidth or increased latency. Our engineering team is actively working on this. We will provide updates as the situation develops. If you have any questions, please reach out to our support team at [email protected] or raise an inquiry in your dedicated Slack channel. Thank You, CoreWeave Support

Hello, We have identified the issue affecting the backend GPU interconnect in US-CENTRAL-08A and mitigation is in progress. Multi-node GPU workloads may still see reduced bandwidth or increased latency while we complete this work. We will update you as the situation develops. We appreciate your patience as we work through this. Thank You, CoreWeave Support

We're investigating degraded performance with metrics in US-EAST. Querying US-EAST metrics in Grafana may be slow or return intermittent errors. Metric ingestion is degraded, but no metric data is being lost, and customer workloads are not affected. We'll provide updates as we remediate.

We have identified the cause of the metrics degradation and are actively working on a fix. We have also taken steps to stabilize the metrics query endpoints, and we continue to monitor overall system health. We will share further updates as remediation progresses.

Hello, We are actively investigating an issue that is impacting our storage services in US-EAST-13A. As the situation evolves, we will continue to update our customers on the matter. If you have any questions, we encourage you to reach out to support by email: [email protected] We apologize for any inconvenience this may cause and thank you for your patience while we work to resolve the matter in a timely manner. Thank you, CoreWeave Support

We have identified the root cause of the issue and are actively working on remediating the cluster. We will update you when we have pushed a fix.

We believe the issue has been mitigated, and we are continuing to monitor the affected services. If you are experiencing an issue or have any further questions please reach out to [email protected].