A live-events company runs a ticketing application on EC2 instances in an Auto Scaling group and uses PostgreSQL for the database. Sale events produce one-time scheduled spikes in demand. The company needs a scaling approach that maximizes availability during these scheduled sale events. Which solution meets these needs?

The correct solution uses scheduled scaling for EC2 instances, which is ideal for predictable, one-time spikes like sale events, ensuring resources are provisioned in advance. Hosting the database on Amazon Aurora PostgreSQL Multi-AZ provides high availability and performance. The key to handling the database load is to use an EventBridge rule to trigger a Lambda function that creates a larger Aurora Replica before the sale. This replica can then be promoted to primary (failover) for the event, handling the increased read and write traffic. After the sale, another EventBridge-triggered Lambda scales the replica back down, optimizing costs. Incorrect options: Aurora Serverless v2 is good for variable loads but doesn't explicitly allow pre-warming with parallel Lambda functions for a scheduled spike, and the replica scaling is automatic, not controlled by a failover to a pre-provisioned larger replica. RDS for PostgreSQL Multi-AZ with automatically scaling read replicas doesn't exist; RDS read replicas are provisioned with a fixed instance size. Failing over to a larger read replica isn't a standard scaling pattern for RDS. Predictive scaling for EC2 is less suitable for one-time, scheduled spikes compared to scheduled scaling, which guarantees capacity. The database pre-warming with parallel Lambdas is not a standard or efficient way to scale an RDS database for a scheduled event.