This epic delivers the full Phase of OSDFM’s Service Cluster autoscaling system. It improves SC upscaling behavior, introduces robust dormant-SC lifecycle management, and strengthens region-level autoscaling controls. The work adds missing metrics and alerts required for new-region readiness, enhances provisioning shard lifecycle handling, standardizes SC creation behavior, and enables autoscaling features across Stage and Production.

1. SC autoscaler logic – SC upscaling behavior, safe SC creation, and saturation calculation.

1. Dormant SC logic added – auto-marking, activation thresholds, and integration with watcher/autoscaler.

1. New-region autoscaling metric improvements – 

added the missing dormant-SC metrics to avoid false alerts, and ensured the Prod dormant-SC-creation alert is triggered only when region saturation reaches 73% and no dormant SC exists. Also added a dormant-SC activation alert when region saturation reaches 90% and no dormant SC has been set to Ready.

1. Provisioning shard lifecycle enhancements – shard-full detection and maintenance handling.

1. Standardized SC creation behavior – first SC created in maintenance, subsequent SCs in ready.

1. Region-level autoscaling opt-out – via sc-autoscale-ignore label.

1. Dormant SC feature enabled in Stage.
2. Dormant SC feature enabled in Production.{}