COSESAI

OpenTelemetry Semantic-Convention v1.20 → v1.27 Attribute Migration Field Guide: When Three Release Layers, One Collector Block, and Four Naive-Regex Failure Modes Decide Whether Your Instrumentation Survives Contact With v1.27

By xiezixing 2026年10月9日

Three semantic-convention releases, twenty-five renames, and a single leftover http.target that no auto-rewriter can split.…

未分类

One Round Trip, Five Bytes: An HTTP/3 QUIC Handshake Anatomy Field Guide

By xiezixing 2026年8月7日

A QUIC handshake is one round trip plus a CRYPTO frame inside it, and the…

未分类

API Latency Budget Planner: Why Most Budgets Are Fiction Until You Compare Them to a Trace

By xiezixing 2026年7月6日

A 300 ms P99 budget that ignores OpenTelemetry measurements is a guess with a flag on it. The three decisions every team makes — cold path, warm path, processing — are almost always the wrong ones until the waterfall meets the trace.

未分类

The Three Numbers That Quietly Decide Your Percentile: Why Your Dashboard Says P99 Is Not What Engineering Calculates

By xiezixing 2026年6月19日

The Three Numbers That Quietly Decide Your Percentile: Why Your Dashboard Says P99 Is Not…

未分类

The Three Environment Files That Hide the Outage You’re About to Ship

By xiezixing 2026年6月16日

The Three Environment Files That Hide the Outage You’re About to Ship Your staging .env…

Cover image showing a YAML config diff between staging and production environments

未分类

The Environment That Wasn’t Like the Other

By xiezixing 2026年6月14日

Most outages that get blamed on "the deploy" are actually config drift. A setting that…

Time Series Anomaly Detector poster: Z-Score and IQR statistical detection methods for monitoring dashboards

未分类

Why Your Monitoring Dashboard Misses the One Spike That Matters

By xiezixing 2026年6月7日

Most anomaly tools chase the same problem: distinguish a real signal from the noise floor…