Starting at 12:28PM UTC, we deployed a new version of our services as we do multiple times a day. Our monitoring tools quickly informed us that it wasn’t going on as planned and we quickly decided to cancel this deployment. 2 out of 6 of our clusters hosted across the globe were impacted. Some technical changes were not executed in the right order on our side and were the cause of the issue.
By 12:44PM UTC, we saw our clusters status going back to normal. We will keep working on preventing that kind of errors from happening. We apologize for the service disruption.