Question

A company runs a serverless REST API using Amazon API Gateway backed by AWS Lambda functions that write to a small Amazon RDS for PostgreSQL Multi-AZ instance. During marketing campaigns, sudden traffic spikes cause thousands of concurrent Lambda invocations to open connections to the database simultaneously, exhausting the connection limit and causing widespread request failures. The team wants to smooth out these bursts so the database is never overwhelmed, while ensuring no write requests are lost. Which solution best meets these requirements?

Accepted Answer

Place an Amazon SQS queue between API Gateway and the processing layer, have API Gateway send writes to the queue, and process messages with a Lambda function whose reserved concurrency limits parallel database connections.

Answer

Increase the Lambda function's memory allocation and enable provisioned concurrency so functions initialize faster during traffic spikes.

Answer

Enable API Gateway caching so repeated write requests are served from the cache instead of reaching the database.

Answer

Migrate the RDS instance to a larger instance class and add multiple read replicas to distribute the connection load.