The service was released which caused a few seconds of downtime. The new zero-downtime deployments are being introduced which will solve this issue.