blainekwilson/failed-request-trace

This project is a diagnostic tool for the NGINX and OpenResty web servers, built to help engineers figure out what happened to a web request that failed or behaved unexpectedly. It was inspired by a similar feature in Microsoft's IIS web server, and it is aimed at modern setups where a request might pass through several layers, such as a load balancer, a firewall, an API gateway, or a container, before reaching the actual application. When something goes wrong, it can be hard to tell where the problem happened. This tool records details about each request and its response, such as which headers were sent and received, whether cookies arrived correctly, and whether something along the way modified the request. It then outputs this information as a structured log entry in JSON format, tagged with a request ID so a specific request can be traced through the system. Security is a stated priority: sensitive information like passwords, authorization headers, and session cookies is automatically hidden in the output by default, though an administrator can choose to reveal specific headers temporarily for deeper troubleshooting. The README shows how to build the tool as a Docker container, run it, and send it a sample request with common troubleshooting headers to see an example trace. Right now it only supports NGINX and OpenResty, but the project lists plans to add support for AWS Lambda, AWS API Gateway, Kubernetes, the Envoy proxy, and Cloudflare Workers in future versions. This is meant for developers and operations engineers who manage web infrastructure and need better visibility into why requests fail, rather than for building an application from scratch.