Services or capabilities described in Amazon Web Services documentation might vary by Region. To see the differences applicable to the China Regions, see Getting Started with Amazon Web Services in China (PDF).

Monitoring Performance

This section describes tools for monitoring an application's performance.

Performance Monitoring using CloudWatch Metrics

You monitor your application's resource usage, throughput, checkpointing, and downtime using CloudWatch metrics. For information about using CloudWatch metrics with your Kinesis Data Analytics application, see Metrics and Dimensions.

Performance Monitoring using CloudWatch Logs and Alarms

You monitor error conditions that could potentially cause performance issues using CloudWatch Logs.

Error conditions appear in log entries as Apache Flink job status changes from the RUNNING status to the FAILED status.

You use CloudWatch alarms to create notifications for performance issues, such as resource use or checkpoint metrics above a safe threshold, or unexpected application status changes.

For information about creating CloudWatch alarms for a Kinesis Data Analytics application, see Alarms.