A global gaming company uses Amazon DynamoDB for storing user profiles. They need to monitor DynamoDB's performance in real-time, specifically tracking read/write capacity utilization, latency, and throttled requests. They also require automated alerts if these metrics exceed predefined thresholds. The solution must be cost-effective and provide granular visibility.
- AImplement custom Lambda functions to periodically poll DynamoDB metrics via API and push them to CloudWatch.
- BExport DynamoDB streams to Amazon Kinesis Data Firehose and then to Amazon OpenSearch Service for real-time dashboards.
- CEnable DynamoDB Contributor Insights to identify frequently accessed or throttled partitions.
- DUse CloudWatch Alarms directly on DynamoDB's built-in metrics to monitor performance and send notifications.
Show answer & explanationAnswer & explanation
Correct answer: D. Use CloudWatch Alarms directly on DynamoDB's built-in metrics to monitor performance and send notifications.
Amazon DynamoDB automatically sends a rich set of performance metrics (like ConsumedReadCapacityUnits, ThrottledRequests, SuccessfulRequestLatency) to Amazon CloudWatch by default. CloudWatch Alarms can be directly configured on these metrics to monitor thresholds and trigger alerts, providing real-time, granular, and cost-effective monitoring without additional services.
Why the other options are wrong
- A. Polling metrics via Lambda is unnecessary and adds cost and complexity, as DynamoDB already publishes these metrics to CloudWatch automatically.
- B. This is an overly complex and expensive solution for monitoring standard DynamoDB performance metrics, which are already available in CloudWatch.
- C. Contributor Insights helps identify hot partitions, but it is not the primary service for overall real-time performance monitoring and alerting on core metrics like capacity utilization and latency.
DynamoDB CloudWatch Metrics
Amazon DynamoDB automatically publishes a comprehensive set of operational and performance metrics to Amazon CloudWatch, offering insights into table and index activity.
- Metrics include read/write capacity, latency, throttled requests.
- Published automatically, no agent or custom code needed.
- Can be used with CloudWatch Alarms for real-time alerts.
- Provides granular visibility into table and index performance.
Memory trick: DynamoDB's 'Built-in CloudWatch' is your instant performance gauge.