All metrics are back to normal, we are marking this issue as resolved.
The issue seems to have been caused by an upstream networking issue. All systems appear to be back online and we continue monitoring the situation.
We are investigating increased latencies with the API.
