Field notes

Writing rooted in CI/CD workflow analytics — measurement pitfalls, telemetry walkthroughs, and recovery clocks that survive real on-call practice.

  • Recovery signals after a bad release

    Mean time to recovery is only as good as the clocks you start when something goes wrong. A look at incident markers that survive contact with real on-call practice.

  • What to ask in a pipeline telemetry walkthrough

    A field checklist for the first hour with a new delivery team: where the logs live, who owns each stage, and which clocks actually tick.

  • Flaky tests and crooked deployment charts

    Unstable suites inflate change-failure rates and train teams to ignore red builds. Here is how to isolate noise before you trust the analytics.

  • Reading lead time without gaming it

    How software delivery teams in Korea and abroad misread commit-to-production clocks — and a practical way to segment the signal before you act on it.