We saw a degradation in API performance starting at around 16.07 UTC, correlating with a lot of incoming traffic to a specific endpoint. We put in place a block for that traffic at 18.33 UTC, after which API performance recovered. We are monitoring the situation.