The issue has been resolved. We are actively monitoring the service.

Core services have stabilised. We are still seeing a degradation on sandboxes and are working on a fix.

Parts of the Prime Intellect platform (sandbox endpoints and related APIs) experienced elevated error rates.