This article demonstrates how to set up end-to-end distributed tracing starting from Amazon CloudFront using OpenTelemetry, enabling organizations to identify performance issues across their entire system.

• CloudFront Functions generate root span trace context and propagate it via request headers
• OpenTelemetry Java Agent automatically instruments demo app running on Amazon EKS
• Lambda function converts CloudFront logs into OpenTelemetry-compatible payloads
• OpenTelemetry Collector receives and processes trace data from multiple sources
• Data Prepper converts OpenTelemetry format to Amazon OpenSearch Service format
• Trace Analytics visualizes request flow and identifies performance bottlenecks
• Uses b3 specification for trace context propagation across services
• Sample code and infrastructure templates available in GitHub repository

This solution provides vendor-neutral, end-to-end observability by tracing user requests from CloudFront through backend services, enabling faster problem identification in distributed systems.