Posted On: Apr 14, 2022

Amazon CloudWatch announces support for additional statistics such as percentiles on CloudWatch Metric Streams. With Metric Streams, you can create a continuous, near real-time stream of metrics to a destination of your choice. Now with these additional statistics you can gain increased visibility on metric patterns and trends, to more effectively monitor the health and performance of your applications.

Metric Streams enable you to keep your Observability tools updated with the most recent metric data from your applications running on Amazon Web Services (AWS). You can use Metric Streams to send metrics to AWS Partner solutions including Datadog, New Relic, Splunk, Dynatrace and Sumo Logic, or to your data lake on AWS, such as Amazon Simple Storage Service (Amazon S3). These additional statistics can help you gain additional insights to measure and maintain your operational goals such as availability or latency. For example, using the new percentile statistics and monitoring the performance of requests handled by your Application Load Balancer using a p98 percentile, helps you to identify the 2 percent of traffic which is above your goal.

To get started with Metric Streams click on the Streams option from the left menu of the CloudWatch console, under the Metrics main menu. You have an option to add additional statistics either to existing or new Metric Streams. Alternatively, you can use the CloudWatch API, AWS SDK, AWS CLI, or AWS CloudFormation to provision and configure Metric Streams.

The additional statistics are now available in all commercial AWS Regions, excluding the Amazon Web Services China (Beijing) Region, operated by Sinnet, and the Amazon Web Services China (Ningxia) Region, operated by NWCD. To learn more about Metric Streams pricing, refer to the Amazon CloudWatch pricing page. To learn more about Metric Streams and these new additional statistics, please refer to our documentation.