Monitoring Internet Performance with Amazon CloudWatch Internet Monitor - Visualizing Availability and Latency
Learn how Internet Monitor tracks application performance over the internet, analyzes availability by ISP, and detects health events.
Overview of Internet Monitor
Internet Monitor leverages internet measurement data from over 500 cities worldwide to analyze traffic from CloudFront distributions and VPCs, monitoring application performance over the internet. It combines internet measurement data collected by AWS globally with your application's traffic data to visualize availability and latency broken down by ISP and region. Unlike traditional synthetic monitoring that measures reachability from probe locations, Internet Monitor detects issues occurring on the actual path of user traffic, providing monitoring closer to real user experience. Creating a monitor takes just a few clicks in the AWS Management Console - simply specify a CloudFront distribution or VPC to start monitoring.
Health Events and Usage Patterns
Internet Monitor generates health events when internet-side issues are affecting your application. It provides information such as availability drops at a specific ISP or latency increases in a particular region. Health events include estimates of affected traffic volume and user count, enabling quantitative assessment of outage impact. You can receive notifications via EventBridge and build automated response workflows that switch CloudFront origins or trigger Route 53 failovers. Visualizing the geographic distribution of traffic is also useful for optimizing CDN edge location selection and regional placement. Health events are classified into two types - availability degradation and latency increase - each with customizable thresholds.
Leveraging Data for Traffic Optimization
Internet Monitor data can be combined with CloudFront origin failover and Route 53 health checks for traffic optimization. When performance degrades for a specific ISP or region, it provides the data needed to make decisions like routing traffic to a different region using Route 53 geolocation routing. You can build workflows that receive health events via EventBridge and use Lambda to automatically modify CloudFront cache behavior or notify operations teams via SNS. Analyzing historical performance data helps identify recurring performance degradation patterns at specific times or in specific regions, informing CDN configuration and origin placement optimization. Enabling S3 log export allows long-term trend analysis and custom report creation with Athena. To gain a deeper understanding of network design with Internet Monitor, specialized books on Amazon are a useful resource.
Internet Monitor Pricing and Configuration
Internet Monitor charges based on the traffic volume of monitored resources (CloudFront distributions, VPCs). You can set the monitored traffic percentage from 1-100%, balancing cost and coverage. For high-traffic workloads, sampling at 10-25% still provides statistically sufficient accuracy. Availability and performance scores are published as CloudWatch metrics, so they integrate seamlessly with existing CloudWatch dashboards and alarms. Adjust health event thresholds to prevent alert fatigue from minor performance fluctuations. There is a limit of several dozen monitors per account, and in multi-account environments, monitors are created individually in each account.
Use Case-Based Deployment Patterns
For global SaaS applications, monitor CloudFront to understand end-user experience by region, identifying underperforming regions and informing decisions on additional edge locations or Origin Shield adoption. For multi-region web applications, use EventBridge + Lambda to automatically adjust Route 53 weighted routing when issues occur with routing to a specific region, diverting traffic away from the affected region. For private applications within VPCs (internal SaaS, remote worker tools), monitor VPC traffic to detect degradation via specific ISPs, providing justification for adding VPN gateways or evaluating a switch to AWS Direct Connect. In latency-sensitive industries such as finance and healthcare, set strict performance score thresholds to detect even slight degradation immediately.
Complementary Use with Synthetic Monitoring
CloudWatch Synthetics (Canary) is synthetic monitoring that periodically sends requests from probe locations to measure availability and response time, complementing Internet Monitor. Synthetics excels at detecting application-side failures (server down, code bugs, certificate expiration), while Internet Monitor detects cases where the application is healthy but issues exist on the path to users. Using both together enables immediate determination of whether a failure is application-originated or internet-path-originated, reducing MTTR (Mean Time To Recovery). Synthetics is strong for transaction monitoring by verifying specific URLs or workflows via scripts, while Internet Monitor excels at broad awareness of ISP/regional outages across a wide area.
Summary
Internet Monitor visualizes application performance over the internet by ISP and region. It automatically detects internet-side outages and notifies you through health events, providing the data needed for traffic optimization decisions when combined with Route 53 geolocation routing and CloudFront origin failover. Combined use with Synthetics accelerates failure isolation, and sampling rate adjustments let you balance cost and coverage.