On January 4, 2022, at 14:41 UTC, we identified an issue where some connections to WebSocket servers were refused due to overloaded servers. We implemented a fix at 15:30 and WS connections returned to normal operation. The total time for the disturbance was From January 3, 2022 11:30 UTC to January 4, 15:30 UTC.
Why did this issue occur?
The issue occurred due to overloaded servers. We quickly scaled up the servers to accommodate the requests.
What will we do to prevent this from happening in the future?
We have already added a monitor that will alert our engineers of any future refused connections so we can take swift action.