Q: You migrate an application from a traditional RDS instance to Aurora Serverless v2. During a sudden 10x traffic spike, the database scales up successfully, but the application crashes heavily citing "Too many connections." Why didn't Aurora solve the connection limits?

Aurora Serverless v2 dynamically scales compute (CPU and RAM) via ACUs (Aurora Capacity Units) in milliseconds. However, it scales the *underlying instance size*. It does not act as a TCP connection multiplexer. When traffic spikes 10x, the application spawns 10x more active TCP connections to the database. Even though the database has the CPU to handle the queries, the raw connection pool limit was breached before the engine could scale up enough to accommodate the new max_connections parameter limit. *Fix:* Serverless databases must always be paired with a connection pooler like Amazon RDS Proxy to efficiently queue and multiplex the massive influx of microservice TCP connections into a small, steady pool of long-lived database connections.