Scaleway 장애 기록

Customers may experience degraded performance on their websites and email services hosted on the pf-012 platform due to high load on an internal component.

Part of our Mistral Small 3.2 infrastructure start becoming un stable with un usual error rates.

The situation is back to nominal state, the root cause has been identified and incident responses are back to normal.

This incident has been resolved.

Due to abnormal temperature, some storage nodes are being shutdown to prevent data loss. This will impact acces to some objects with one-zone storage class

Servers have been moved to a different location and brought back online. All objects, including one-zone objects should be accessible.

Customers may experience delays when provisioning new Managed Database instances or performing operations that require infrastructure changes in the fr-par region due to an issue with an internal dependency.

This incident has been resolved.

Between 11:50 a.m. and 1:30 p.m., metrics ingestion on FR-PAR was affected due to an operation performed on one of the components to improve its resilience. Metrics and log data are also affected.

We are continuing to monitor for any further issues.

The issue has been identified and a fix is being implemented.

Following a component overload in Cockpit during an upgrade designed to improve its resilience, difficulties integrating logs and metrics arose from Scaleway products between 11.30 am and 3.30 pm. There may have been a partial loss of this data.

Customers may experience missing monitoring data for NATS, Queues, Topics & Events in the fr-par region.

The issue has been resolved and monitoring data is now being collected as expected.

Customers may iSCSI connections failures on Dedibox RPN-SAN HA HDD-18.

Customers may experience iSCSI connection failures on Dedibox RPN-SAN HA HDD-18 in DC2/DC3. The issue has been identified and is currently being monitored. We apologize for any inconvenience caused and will provide further updates as soon as possible.

Situation is now stable, we are closely monitoring situation on both members of HA pair.

Customers may experience partial outages when managing Kafka resources due to an issue on the control plane.

This incident has been resolved.

Customers may experience a partial outage of public network connectivity on Elastic Metal servers in PAR2 (DC5).

The root cause has been found. A fix is being deployed.

The fix has been deployed. Services are up and running. We are monitoring the situation and we will provide more information.

During routine maintenance, a new configuration was deployed to our backbone routers managing traffic for fr-par-2. An unforeseen configuration conflict resulted in a total loss of public connectivity (blackholing) for all Elastic Metal servers. The configuration was pushed at 08:42 UTC for both IPv4 and IPv6. IPv4 was rollback at 09:20 and IPv6 was rollback at 09:45, restoring all connectivity.

Everything is back to normal. Services are up and running. If you are still encountering an issue, please contact our support team.

Today, from 04h48 to 04h50 UTC a block cluster hosting b_ssd volumes reported high latencies following a network issue affecting a server rack Cluster recovered at 04h51. Very few IO operations were still stuck until manual intervention at 05h10 UTC. Situation is normal since.

Customers may be experiencing boot issues after installing / rebooting their Elastic Metal server on PAR2. We are currently investigating.

The issue is not limited to PAR2; it affects all regions on various operating systems. We are still investigating.

We are continuing to investigate this issue.

This incident has been resolved.

Customers may experience connectivity issues with servers in rack B61 due to an unreachable switch in the network infrastructure.

Customers may experience instabilities when using the mistral-medium-3.5-128b model in the Generative APIs service in the fr-par region.

Customers may experience instabilities when using the mistral-medium-3.5-128b model in the Generative APIs service in the fr-par region.

We are working on a fix. Instabilities may still appear when using mistral-medium-3.5-128b model. But overall the service remain usable.

Customers may experience connectivity issues on their VPS instances due to a networking issue on an internal component in the fr-par-1 region.

This incident has been resolved.

Some customers are experiencing errors when using Grafana.

The issue has been identified and a fix is being implemented.

A fix have been deployed and we are monitoring the situation.

This incident has been resolved.

Customers may experience instabilities when using the mistral-medium-3.5-128b model in the fr-par region.

This incident has been resolved.

Customers may experience failures when creating dedicated deployments on the Generative APIs product in the fr-par region.

We are continuing to investigate this issue.

Customers may experience failures when creating dedicated deployments on the Generative APIs product in the fr-par region. The issue has been identified and our teams are working on a resolution.

A fix has been implemented and we are monitoring the results.

This incident has been resolved.

A small subset of Scaleway Product logs/metrics might be unavailable on fr-par on small time-range Yesterday : 9:20 - 11h30 CEST Today : 17:20 - ongoing CEST

We are continuing to investigate this issue.

Observability volumes are increasing, and we are reaching some limits with serverless log ingestion. We have resized the cache for our ingestion components, and we plan to deploy smart features to this component this week.

Some users may experience connectivity issues when accessing newly created resources in VPC. The incident started at 08:50 UTC. The problem has been identified, and we are actively working to resolve it.

We have resolved a certificate configuration issue that caused a communication outage between two of our internal components. All affected services are now fully restored and operating normally. Our team is actively monitoring the situation and implementing permanent changes to improve stability and prevent this specific issue from recurring. The incident was resolved at 10:47 UTC.

After an extended period of monitoring, we have confirmed that all systems are stable and fully operational.