Freshness pressure Bookings is stale for EMEA and APAC at 6:55 a.m. The global dashboard opens in 35 minutes, and the orchestrator only says that the final model succeeded. A green final job can still publish stale source data. The decision is about safe consumer use, not only pipeline status. Incident walk Impact -> Boundary -> Mitigation Manage freshness by moving from the business decision to the failed technical boundary, then to the smallest reversible consumer-facing mitigation. Reflex Rerun everything or wait silently. Users get a truthful state while engineers recover the pipeline. Freshness is a promise to consumers, not…
Read the full lesson
Sign up free: one personalized lesson every day, matched to your role and goals.
Already have an account? Sign in