Skai Outage History

Our engineering team has continued investigating the delay affecting our internal data synchronization processes. We’ve now identified the issue as originating within our own systems rather than with an external provider, as initially believed. The issue has been isolated and remediation is currently underway. We currently expect full recovery within approximately 2 hours. Once recovery is complete, affected processes and features will resume normal operation, and we will also handle any impacted portfolios immediately. We apologize for the disruption and appreciate your patience while we complete recovery efforts

Looks like Microsoft started recovering and we see reports coming in to our platform. There may still be delays in features that rely on yesterday's data until all of yesterday's data will be fetched and processed. As for optimization, no need to move to manual bidding, Skai platform will recover automatically.

The issue was resoled by Microsoft and the reports are coming in. We will continue to monitor.

AWS experienced a major outage at around 3 AM EST, centered on the US-East-1 region. The disruption impacted a wide range of major apps, websites, and services globally. The outage lasted roughly three hours, with most services now reported as fully mitigated. AWS Status page: https://health.aws.amazon.com/health/status Our platform is fully operational for all clients. Performance data has already begun to recover automatically, and any processes waiting for daily data will resume as soon as the data becomes available. Our teams continue to monitor the situation closely to ensure there is no further impact.

Amazon continues to work on mitigating the issue on their end. You can monitor their official status page for updates: https://health.aws.amazon.com/health/status We are actively monitoring the situation while awaiting a full resolution from Amazon. We are still experiencing delays in Amazon Performance data being received due to the ongoing incident. Data is continuing to process, albeit more slowly than usual. Some users may notice slower performance when loading analysis grids; this can often be resolved by refreshing the page. A few users have encountered intermittent “Internal 500” errors. If this occurs, please try again after a short interval. We hope Amazon will resolve the underlying issue soon to restore normal service levels.

We are continuing to monitor for any further issues.

Amazon AWS has confirmed that the issue has been resolved on their side. For reference, see the AWS Status page: https://health.aws.amazon.com/health/status A backlog of Amazon performance reports remains on our side. We are actively monitoring and processing these reports to ensure normal operations are fully restored.

Amazon has confirmed that the earlier issue has been resolved. However, we are still seeing delays in Amazon Performance data being received. Our engineering team is actively working with Amazon to investigate and expedite a full resolution. We appreciate your patience and understanding while this is being addressed.

We can confirm that Amazon Performance data has now fully recovered, and all services are fully operational for all clients. Our team has verified that there is no further impact across processes, including: - Scheduled Reporting - Automated Actions - Optimization Thank you for your patience while this issue was being resolved.

Daily reports have been processed successfully. We will continue to monitor.

Criteo informed us that the issue was fixed, and we do see an improvement in our API calls to Criteo, we are still monitoring the situation

Our API calls go through to Criteo, and we do see a decrease in our requests queue. However, due to Criteo's API call limit, the decrease is slow. We hope to get it all done until the beginning of the business day tomorrow (US). Operations that rely on performance data availability may still have degraded performance, as performance data us still being fetched, limited by Criteo's API call limit. Uploads and Downloads work with no issue.