Replayable Automation Needs Better Run Receipts

Описание и применение

Автоматизация становится более прозрачной и отлаживаемой благодаря компактным JSON‑получениям, которые фиксируют идентификацию, входы, решения, попытки, доказательства и очистку, позволяя быстро находить и воспроизводить ошибки.

Задача

Automation failures are hard to diagnose because run logs lack concise, actionable context, making it difficult to understand why a job failed or how to reproduce the issue.

Результат

Automations become more transparent and debuggable; developers can quickly locate a run, understand every decision, and replay the scenario with controlled fixtures, reducing mystery and improving reliability.

Рабочий процесс

• 1. Execute automation job with unique run ID and commit SHA.
• 2. Capture safe inputs (feature flags, endpoint names).
• 3. Log decisions (continue, retry, stop) and attempt details (start/end time, status, correlation ID).
• 4. Record evidence (response status, headers, redacted IDs).
• 5. Note cleanup actions (data removed or left).
• 6. Store the receipt as a JSON artifact in CI logs.
• 7. Use the receipt to replay decisions with safe fixtures and compare outcomes.

Доказательства

A receipt turns an automation run from a single outcome into something a developer can replay and understand.

Источник

Start with one JSON file per run. Add richer tracing only when the receipt shows a real gap.

Источник