CloudWatch Flow Monitors: RTT scale issue ?

Thank you for bringing this to our attention. You're correct in noticing a discrepancy between the CloudWatch Flow Monitor's reported Round-Trip Time (RTT) and what you're observing with a simple ping test.

The CloudWatch Network Monitoring feature, which includes flow monitors, is designed to provide near real-time visibility into network performance for AWS workloads. However, the scale issue you're experiencing with the RTT metric is not typical and may indicate a potential problem with the metric calculation or display.

The RTT metric in CloudWatch Network Monitoring is indeed supposed to be measured in milliseconds. It represents the average round-trip time for network flows monitored by your flow monitor. This metric should typically align more closely with what you observe using other network diagnostic tools like ping, especially for traffic between instances in the same region.

A difference of 2000 ms (2 seconds) compared to 2 ms is significant and suggests there might be an issue with how the metric is being calculated or displayed in your specific case. This could be due to several factors:

1. A bug in the CloudWatch Network Monitoring service for your region or account.
2. Misconfiguration of the flow monitor.
3. An issue with how the data is being aggregated or displayed in the CloudWatch console.

To troubleshoot this issue:

1. Double-check your flow monitor configuration to ensure it's set up correctly.
2. Try creating a new flow monitor with the same settings to see if the issue persists.
3. Check if the problem occurs for other metrics provided by the flow monitor, such as data transferred or retransmissions.
4. Consider monitoring the RTT over a longer period to see if there are any fluctuations or if the high values are consistent.

If the issue persists, it would be advisable to contact AWS Support for further investigation. They can look into the specifics of your account and region to determine if there's a known issue or if this is an isolated incident that needs to be addressed.

In the meantime, you can continue to rely on other network diagnostic tools like ping for accurate RTT measurements between your instances. The CloudWatch Network Monitoring should provide valuable insights, but in this case, the RTT metric appears to be inaccurate for your specific setup.
Sources
Amazon CloudWatch adds network performance monitoring for AWS workloads using flow monitors - AWS
View Network Flow Monitor metrics in CloudWatch - Amazon CloudWatch