SLOs & Data Incidents — Week 4 CheckWeek 4 of Data Quality, Contracts & Observability: SLIs, error budgets, and incident practice for data — from the SRE borrowing to Agoda's financial-metrics platform.

Why is a deliberately sub-100% SLO target the MECHANISM rather than lowered ambition?

Which TWO SLIs are specifically motivated by data's nastiest failure class — the silent error?

A table promises 'lands by 06:00, 99% of days' over a 90-day quarter. At most how many late landings keep the promise?

Why is an alert that fires AT the landing deadline nearly useless for meeting the SLO?

Agoda published 95.6% uptime against a 99.5% target for its financial-metrics platform. Why is publishing that GAP a sign of a healthy SLO regime?

The hardest part of Agoda's consolidation of financial pipelines wasn't Spark. What was it?

Expedia built its own SLO platform on top of Datadog. What was the BINDING constraint that shaped the architecture?

A data incident is confirmed at 07:10; the dashboard's consumers typically look at 09:00. What does mature incident practice do at 07:15?