Some deployments fail to start and/or are not being properly reported by the API. We are investigating.
08:38 UTC: We are restarting part of the deployment system.
08:49 UTC: Since 5 minutes ago, deployments are being processed with some delay.
08:54 UTC: Back to normal.