Initialize Network Flow Monitor

Before you can view performance metrics for network flows, you must initialize Network Flow Monitor, which grants required permissions and creates an initial topology for your account or accounts. If you plan to monitor resources for multiple accounts, you must also configure AWS Organizations with Amazon CloudWatch. Then, you specify accounts for your Network Flow Monitor scope, so that Network Flow Monitor can create an initial topology for all the accounts that you'll be tracking performance metrics for.

In addition, you must install agents on your instances, to send performance metrics to the Network Flow Monitor backup ingestion server. For more information, see Install Network Flow Monitor agents on instances.

The steps that you take to initialize Network Flow Monitor vary depending on whether you are measuring performance metrics for resources in a single account, or you want to monitor metrics from resources that are owned by multiple accounts in your organization.

Install agents on instances

To track network performance with Network Flow Monitor, you must initialize the service, but you must also install Network Flow Monitor agents on your workload's EC2 instances and add permissions for the agents to send networking performance metrics to Network Flow Monitor. After you install the agents, wait a short period of time (about 20 minutes), for data to begin being sent to the Network Flow Monitor backend. Then, you can view network performance metrics, on the Workload insights tab, and also create monitors, to view detailed information.

For example, you can view the top contributor performance metrics for data transferred and retransmission timeouts, for network flows between your local and remote resources, collected by Network Flow Monitor agents. By viewing and analyzing these metrics, you can choose specific flows that you want to see more details for and track more closely with a monitor. By creating a monitor for specific flows, you can see detailed information about them, including metrics, sorted by the top contributors for each metric type and topologies for each network flow.

With a monitor, Network Flow Monitor also provides a network health indicator (NHI), which you can use to see if there have been AWS network impairments for network flows that you're tracking in the monitor, during a time period that you've selected. That information can help you decide where to focus your network troubleshooting efforts.

For more information, and instructions for how to install agents, see Install Network Flow Monitor agents on instances.