Question

A company runs a mission-critical DynamoDB table with provisioned capacity. During unpredictable traffic spikes, the application experiences ReadThrottleEvents that cause API timeouts. The DevOps team wants an automated response that increases the table's provisioned read capacity when sustained throttling occurs, but only after confirming the condition persists for several minutes to avoid reacting to brief bursts. The solution must require no standing infrastructure and must log every remediation action for auditing. Which approach best meets these requirements?

Accepted Answer

Create a CloudWatch alarm on the ReadThrottleEvents metric with an evaluation period spanning several minutes; on ALARM, trigger an EventBridge rule that invokes a Lambda function to update the table's provisioned capacity, and write structured log entries to CloudWatch Logs.

Answer

Enable DynamoDB auto scaling with a target utilization, and configure an EventBridge rule on the scaling activity to notify the team via SNS.

Answer

Create an EventBridge rule that matches every ReadThrottleEvents data point emitted to CloudWatch and directly invokes a Systems Manager Automation runbook to raise capacity.

Answer

Schedule a Lambda function every minute using EventBridge Scheduler to poll the ReadThrottleEvents metric with GetMetricData and adjust capacity when any throttling is detected.