Posted On: Dec 10, 2020
Amazon EC2 now provides additional network performance metrics to help customers gain more insights into instance network performance. Five new metrics provide customers visibility when their instances exceed network allowances defined by AWS. This visibility helps customers proactively resolve application performance issues and right size instance fleets based on desired network performance.
Prior to this announcement, customers had the capability to monitor network bytes and packets utilization on EC2 instance. Amazon EC2 defines network allowances on an instance level to ensure a high quality customer networking experience, including fairness between customers and consistency of networking performance across instance sizes. With this launch, customers can now monitor when AWS allowances are exceeded to troubleshoot and analyze root cause for network performance issues. The new metrics inform customers in real time of network traffic impacted when instance allowances for inbound and outbound bandwidth, packets-per-second (PPS), connections tracked and PPS to link-local services are exceeded. Utilizing these metrics, customers can quickly troubleshoot application performance issues related to exceeded allowances and proactively scale-up their instance fleets to accommodate a surge in network traffic. These metrics will also help customers right-size their EC2 instances from network standpoint allowing them to run benchmark tests before production deployment.
These metrics are available on instances running Linux ENA driver starting from version 2.2.10, Windows ENA driver version 2.2.2, along with DPDK ENA driver version 20.11 and FreeBSD ENA driver version 2.3.0. They can be accessed from within the instance at no extra cost using simple command line tools. Customers can download corresponding ENA driver from github or use latest AL2 AMI builds. Latest version of the CloudWatch agent will support these as custom metrics for trending and alerts.
Instance Level Network Performance Metrics is available in all AWS Commercial and GovCloud (US) Regions, with the exception of China (Beijing) and China (Ningxia).
To get started on these metrics, please see below user guides:
Monitoring Network Performance Metrics for Linux
Collecting Network Performance Metrics with CloudWatch
Blog on Network Performance Metrics