Visualizing Edge Application Performance with Amazon QuickSight Sankey Diagrams

In modern gaming and edge computing, latency is the silent killer. When players in Tokyo experience lag connecting to a server in Singapore, time-series line graphs often fail to show where the bottleneck is. Is it the regional gateway? The edge node? Or the core database?

This tutorial demonstrates how to build a Network Performance Pipeline for a fictional gaming company, LatencyLabs. We will visualize traffic flows and latency between distributed nodes using Amazon QuickSight Sankey Diagrams, powered by Athena and S3.

The Architecture

Step 1: Athena Table Configuration

First, we tell Athena how to interpret our S3 CSV logs.

CREATE DATABASE network_performance;

CREATE EXTERNAL TABLE network_performance.ping_logs (
    timestamp_utc STRING,
    protocol STRING,
    source_ip STRING,
    source_region STRING, -- e.g., 'us-east-1'
    dest_ip STRING,
    dest_region STRING,   -- e.g., 'ap-northeast-1'
    rtt_msec DOUBLE
)
ROW FORMAT DELIMITED 
FIELDS TERMINATED BY ',' 
LOCATION 's3://latencylabs-logs/ping-data/';

Step 2: Configuring QuickSight

Once the data is in Athena, setting up QuickSight is a drag-and-drop process.

• Data Source: Select "Athena" and choose the network_performance database.
• Visual Type: Click the "Sankey Diagram" icon in the Visuals pane.
• Field Mapping:
  • Source: Drag source_region here.
  • Destination: Drag dest_region here.
  • Weight: Drag rtt_msec here.

Result: You will see thick lines connecting regions with high latency (high RTT), and thin lines for fast connections. High latency paths immediately stand out visually as "fat pipes" of delay.

Conclusion

By combining S3, Athena, and QuickSight, LatencyLabs created a "Mission Control" dashboard that visualizes global network health at a glance. What used to be a spreadsheet of 10,000 latency rows is now a single, interactive diagram that highlights exactly where the lag is coming from.