Dither and Bicker

Pragmatic software engineering notes on web telemetry, search indexation, SQL internals, and QA automation. Written by Chris since 2010.

RIP Google Analytics, Casualty of Spam (And What Replaced It)

How Measurement Protocol referral spam polluted client-side analytics across the open web, why Ghost Referrers broke attribution for independent publishers, and why raw server edge logs and direct search index verification remain the only ground truth.

Read full article →

Diagnosing "Crawled — Currently Not Indexed" Across Large Content Architectures

A systems-level look at how search engine schedulers prioritize crawl queues, why thin render trees get dropped after initial fetch, and how engineering teams audit real-time URL indexation status without triggering rate limits.

Read full article →

Deterministic Email OTP and Webhook Testing in Autonomous CI/CD Pipelines

Why flaky end-to-end authentication tests stall deployment pipelines, and how catch-all disposable test inboxes, webhook inspectors, and Model Context Protocol (MCP) endpoints let coding agents verify signup flows in milliseconds.

Read full article →

Using ETag and If-Modified-Since to Cut Origin Bandwidth and Crawl Waste

Practical notes on HTTP conditional requests, weak vs. strong ETags behind compression proxies, and how 304 Not Modified responses save bandwidth for crawlers and clients alike.

Read full article →

PostgreSQL Partial Indexes for Sparse Job Queues and Soft-Deleted Tables

How attaching a WHERE predicate to a PostgreSQL index shrinks B-tree memory footprints by orders of magnitude when 99% of rows are already processed.

Read full article →

Structured Edge Access Logging Without Accidental PII Leakage

How to scrub sensitive query parameters, authorization tokens, and raw client IP addresses at the edge before shipping HTTP access telemetry to long-term storage.

Read full article →

DNS TTL Tradeoffs During Zero-Downtime Infrastructure Migrations

Why lowering DNS TTL 48 hours before a server or CDN cutover prevents split-brain traffic and spurious TLS or 5xx errors while resolvers converge.

Read full article →

MySQL: Finding Records in One Table That Are Not in Another Table

Benchmarking LEFT JOIN ... WHERE IS NULL against NOT EXISTS and NOT IN subqueries across millions of rows in InnoDB, and how nullable foreign keys silently break naive anti-joins.

Read full article →