Getting insights to improve application performance in Amazon CloudWatch Internet Monitor (Traffic insights tab) - Amazon CloudWatch
Services or capabilities described in Amazon Web Services documentation might vary by Region. To see the differences applicable to the China Regions, see Getting Started with Amazon Web Services in China (PDF).

Getting insights to improve application performance in Amazon CloudWatch Internet Monitor (Traffic insights tab)

Use the Traffic insights tab in the CloudWatch console, under Internet Monitor, to look at summary information for top traffic (by volume) for your application. You can filter and sort your application traffic in multiple ways. Then, scroll down, and select different setup combinations for your application to see what Internet Monitor suggests for the best alternatives to obtain the fastest time to first byte (TTFB) performance.

Internet Monitor publishes to CloudWatch Logs internet measurements every five minutes for the top 500 (by traffic volume) city-networks (that is, client locations and ASNs, typically internet service providers or ISPs) that send traffic to each monitor. Optionally, you can choose to publish internet measurements for all monitored city-networks (up to the 500,000 city-networks service limit) to an Amazon S3 bucket. For more information, see Publishing internet measurements to Amazon S3 in Amazon CloudWatch Internet Monitor.

Top traffic summaries

You can start by viewing high-level summaries of your application's overall traffic and performance, over a specific time period, filtered by client location. You can also look at performance for your application for the top (or bottom) client locations by traffic volume, filtered and sorted in multiple ways. For example, you can sort by granularity (that is, city, subdivision, country, or metro area), by total traffic, average time to first byte (TTFB), and other factors.

To learn more about client location accuracy in Internet Monitor, see Geolocation information and accuracy in Internet Monitor.

Note

The filters that you use apply to whole page, so they affect which city-networks are included in the summary graphs and information for total traffic, and also which city-networks are included in the Traffic optimization suggestions section that follows.

Traffic optimization suggestions

The Traffic optimization suggestions section displays a filtered set of monitored city-networks (locations and ASNs, internet service providers) for your traffic, along with the total client traffic for each one. The entries in the table is based on the filters that you chose for your application traffic for Traffic insights at the top of the page. The default is the top 10 cities by traffic volume. You typically see more than 10 rows in the table, because there's entry for each unique city-network pair. That is, there's one row for each combination of location (city) and ASN (network provider) that clients access your application through—such as, Dallas, Texas, USA and Comcast, for example.

Note

To see traffic optimization suggestions for all your monitored city-networks, you can run a query directly in CloudWatch Insights. For a sample query that doesn't include the geographical granularity filter that limits the city-networks list on this page, see Using CloudWatch Logs Insights with Amazon CloudWatch Internet Monitor.

In this section, select different options: Amazon EC2, CloudFront, or both. This lets you see what the predicted average time to first byte (TTFB) values are for clients when you use your application with those services in different Amazon Regions, compared to the current TTFB. For more information about TTFB calculations, see Amazon calculations for TTFB and latency.

By selecting different the options, and then viewing the results in the table, you can start planning setups and deployments that can improve performance for your clients. Note that you might see a dash (-) instead of a value in a column, when data is not available to display. To review a specific example of how to improve performance, see Using Amazon CloudWatch Internet Monitor for a Better Gaming Experience.

For example, to get started, for a specific city-network (client location and ASN pair), experiment with selecting either the EC2 or CloudFront option, or both. For each city-network listed in the table, Internet Monitor shows you the potential performance improvements to the TTFB, based on a traffic routing choice (through a specific Amazon Web Services Region) with that option, compared to the current setup. (Note that, for completeness, the table also includes routes that are already optimized.) For example, you might see a predicted average TTFB that is 50ms for using EC2 routing through us-east-1 compared to your current setup with a TTFB of 100ms where you're using EC2 routing through us-west-2. So you might consider routing through us-west-2.

As another example, you might select EC2, and then see that it doesn't make a measurable performance difference for one client location and ASN, but then note that when you select CloudFront with the same Region, it lowers the TTFB somewhat. This suggests that if you add a CloudFront distribution in front of your application, it could result in a performance improvement and might be worth trying, for this client location and ASN.