Evaluate network flows with workload insights - Amazon CloudWatch

Evaluate network flows with workload insights

Network Flow Monitor provides workload insights about the network flows in the scope that you enable monitoring for. By surfacing top contributor network flows for each metric type, you can see which flows are potentially experiencing issues. You can view these top contributor metrics in the console on the Workload insights tab. In Network Flow Monitor, top contributors are network flows that have the highest values for each network performance metric.

Top contributors

On the Workload insights page in the Network Flow Monitor console, Network Flow Monitor displays the top contributor network performance statistics for the network flows between all the resources in your monitoring scope where you've deployed agents.

To compile lists of top contributors, Network Flow Monitor determines the network flows in your scope that have the highest values for retransmissions, retransmission timeouts, and data transferred. These network flows are the top contributors for each metric type.

For workload insights, the top contributors are determined for all the network flows that you're receiving performance information for; that is, network flows for all of the resources in your scope with Network Flow Monitor agents installed.

Network Flow Monitor classifies metrics into designated destination categories. Metrics are only shown in Workload insights for traffic flows in these categories. Metrics are displayed for the following categories of flows:

  • Between AZs (INTER_AZ)

  • Within AZs (INTRA_AZ)

  • Between VPCs (INTER_VPC)

  • Between Regions (INTER_REGION, which means performance for network flows between resources in your Region and the edge of another Region)

  • Toward Amazon S3 buckets (AMAZON_S3)

  • Toward Amazon DynamoDB (AMAZON_DYNAMODB)

Flows that can't be classified into one of these destination categories are included in the UNCLASSIFIED category. For example, traffic that goes to the internet is included in the UNCLASSIFIED category. Unclassified flows are not included in Workload insights.

Performance metrics

Performance metrics on Workload insights are shown in separate tables for the following metric type: Retransmissions, retransmission timeouts, and data transferred. The data provided is for the top contributors for each type. Note that after you first install Network Flow Monitor agents, there is a waiting period (about 20 minutes) before you can view performance metrics, while agents gather and send data to the Network Flow Monitor backend.

Monitor specific network flows

As you review performance metrics, when you see specific resources or network flows that you want to explore more details for, you can create a monitor that includes just those flows.

With a monitor, you can track specific groups of network flows over a period of time, to gain insights into one aspect of your workload. You can also get helpful troubleshooting information, such as checking the network health indicator (NHI) to see if issues you see are caused by AWS impairments.

For more information, see Create and work with monitors in Network Flow Monitor

Multi-account coverage

To work with multiple accounts in Network Flow Monitor, you must configure AWS Organizations integration with CloudWatch. By configuring Organizations, you can add accounts to the scope for your Network Flow Monitor coverage. Then, if you have multiple accounts in your scope that each have resources that you want to monitor network flows between, you can specify a scope that includes all of the accounts, and then view performance metrics gathered by the Network Flow Monitor agents that you've installed on your resources. For more information, see Initialize Network Flow Monitor for multi-account monitoring.