No components marked as affected
Resolved
The incident was resolved on January 11, at 16:00 UTC.
Identified
Cognite engineering has re-opened the incident from the 3rd of January after observing that we continue to have elevated error rates for the EUR-W3 cluster. The user experience is that we return 5xx errors on requests to RAW and that transformations jobs using RAW are failing or timing out. The engineering team is working on improving the performance of the underlying database and is also investigating the type of requests that dominates the traffic during periods with elevated error rates.