A slowdown of our primary BigTable cluster was causing some scoring requests to time out and a HTTP 500 to be returned. We have now switched to our secondary BigTable cluster which is fine, and the problem has been immediately resolved.
Posted Feb 08, 2023 - 16:48 GMT
We're experiencing an elevated level of API errors and engineers are currently investigating the issue. Please check back here for further updates.
So far, we have identified that the requests affected are scoring requests, potentially related to increased database latencies.