Production Observability at System Boundaries

Field notes on proving telemetry across proxies, runtimes, queues, edge controls, and production logs without confusing configuration with evidence.

Why Configured Telemetry Does Not Prove Production Tracing

Production tracing records state transitions but cannot make queue commits atomic or prove that retries and side effects are correct.

Preserving the Real Client IP Through a Proxy Chain That Rewrites the Evidence

Protect client IP provenance across proxy chains by validating peers, enforcing provider contracts, and overwriting one canonical header.

Replacing a Disappearing Sentry With an End-to-End Datadog Observability Plane

Replace unverifiable parity with a tested investigation across Cloudflare and Datadog, including joins, logs, ownership, and proof limits.

Privacy-Safe Caddy Access Logs for Production Debugging

Build access logging with privacy-safe Caddy records, one-event retrieval, observed retention, and deliberate loss detection.

How to Harden Cloudflare WAF Without Bot Management

Keep Cloudflare WAF enforcement scope within reviewed evidence using narrow rules, conservative sampling, phase checks, and rollback.

Measuring Performance Work Before Declaring Victory

Use performance attribution to compare old and new paths, calculate database deltas, control confounders, and decide what to retain.

Measure Django Cache Performance Before Adding More Caching

Measure Django cache performance by matching requests, tracing get, loader, and set stages, and testing whether hits remove limiting work.