Scheduled Scaling Policies for ECS Fargate Batch Processing

Act as a Systems Automation Engineer. Configure a highly efficient, cost-saving Scheduled Scaling architecture specifically for an ECS Fargate service named [SERVICE_NAME] that exclusively processes heavy, nightly batch computation jobs. Provide the exact AWS CLI `put-scheduled-action` commands required to physically scale the service up to a massive [DESIRED_COUNT] tasks utilizing the precise [CRON_EXPRESSION_START]. Subsequently, configure a secondary trigger utilizing [CRON_EXPRESSION_STOP] to violently scale the desired count back to absolute zero. Crucially, outline the programmatic protection logic required to intercept the scale-down event and prevent Fargate from killing tasks that are mid-computation, ensuring data integrity for unfinished batch jobs.