A Lambda function services read-only queries to an Amazon Aurora MySQL cluster. EventBridge events invoke the Lambda, and as invocation rate rises, database latency increases and throughput drops. Which combination of changes will improve performance? (Choose three.)

Database connection overhead is a common cause of performance issues with Lambda. Opening a new database connection for each invocation is inefficient and can overwhelm the database. Provisioning an Amazon RDS Proxy connected to the Aurora cluster reader endpoint and having the Lambda function connect to this proxy is crucial. RDS Proxy manages connection pooling, reducing the number of direct connections to the database and improving efficiency. Setting a maximum connections percentage on the proxy prevents it from overwhelming the database. Opening the database connection outside the Lambda event handler (global scope) allows the connection to be reused across multiple invocations of the same Lambda instance, significantly reducing connection overhead and improving performance. Implementing connection pooling inside the Lambda code is redundant when using RDS Proxy, as the proxy handles this. Opening and closing the database connection inside the Lambda event handler for each invocation is inefficient and causes the problem. Connecting directly to the Aurora cluster endpoint bypasses the benefits of RDS Proxy's connection management.