The original title is: "Debug a Legacy Frontend-Backend Deployment With One Trace ID"

Short summary

A practical runbook for debugging layered legacy deployments by tracing a single request ID across DNS, CDN, frontend, backend, and database layers. Inject revision markers into artifacts and health endpoints, generate request IDs at the edge, and build an evidence table per layer to isolate which component actually failed. After recovery, reproduce in staging and add the missing invariant to your pipeline.

• •Trace one request ID across all layers instead of upgrading everything blindly
• •Record artifact digests, source revisions, and config versions at deploy time for provenance
• •Build a per-layer evidence table to distinguish stale assets, bad routes, app errors, and DB saturation

Generated with AI, which can make mistakes.