Scaleway Outage History

Scaleway's regional API in Milan became unreachable from Paris, causing Grafana data sources to fail syncing and preventing users from listing their monitoring data sources. This impacted users' ability to manage and visualize their monitoring data through the Cockpit service. The incident lasted 10 minutes and was classified as minor.

Scaleway's payment service provider processed duplicate SEPA payments for some clients, affecting the Excellence component. The company worked with their payment provider to cancel the erroneous duplicate charges. The incident was resolved after approximately 67 hours.

A public network switch failed in Scaleway's DC3 datacenter room 46-D2 starting at 02:02AM, impacting connectivity for some services in that location. The technical team replaced the faulty switch and monitored the replacement to ensure stability. The incident was fully resolved after 2.8 hours of downtime.

Scaleway's Generative APIs experienced a partial outage affecting the gpt-oss-120b and mistral-small-3.2-24b-instruct-2506 models, causing failures when users attempted to use these specific AI models. The incident lasted approximately 14 days before being resolved. The outage was classified as minor and only impacted these two particular models within the generative AI service.

Scaleway's Cockpit service in the fr-par region experienced elevated error rates on custom metrics data ingestion for some customers over 17.7 hours. The issue was caused by a constant increase in the number of series from Serverless services overwhelming the ingestion system. The incident was resolved with monitoring in place for a long-term fix to prevent recurrence.

Scaleway's FR-PAR-2 zone ran out of available public IPv4 addresses, preventing users from creating new instances that required public IPv4 connectivity. The IPv4 stock shortage affected the Instances service in the fr-par-2 region for approximately 1 hour before being resolved.

Scaleway's Cockpit service experienced partial errors in metrics ingestion for Custom and SCW metrics across the fr-par-1, fr-par-2, and fr-par-3 regions starting at 1:00 CEST. The incident affected the ability to properly ingest metrics data, though the service's retry mechanism was expected to prevent data loss. The issue was resolved after 23.8 hours of investigation and remediation efforts.

Scaleway's Cockpit service experienced performance issues with custom data source read/write operations for metrics and logs in the fr-par region starting March 9, 2026. The incident affected the fr-par-1 and fr-par-2 components and lasted approximately 47 hours. The issue was resolved on March 12, 2026.

An internal issue prevented users from creating new Kapsule clusters in Scaleway's Paris region (fr-par) for 43 minutes. The engineering team identified the root cause and implemented a fix, restoring full cluster creation functionality. The incident was classified as minor severity and only affected new cluster provisioning, not existing clusters.

Scaleway's Edge Services experienced a partial outage where GET requests randomly returned truncated HTTP content when fetching from origins, caused by a software upgrade. The issue affected content delivery for approximately 30 days before being resolved through a rollback of the problematic software upgrade.