foutbudget

Foutbudget, in English commonly called an error budget, is a metric used in reliability engineering to quantify the amount of unreliability a system may exhibit without violating its service level objectives (SLOs). It defines the permissible level of failures, latency violations, or other degradations within a specified time window.

It is calculated as the portion of time or requests that may be degraded while still meeting the SLO, effectively the difference between the target reliability and actual performance. For example, with an SLO of 99.9% availability over a 30-day month, the error budget equals 0.1% of the time in that window, roughly 43 minutes of downtime per month. For latency-focused SLOs, it represents the amount of latency or error fraction that may be tolerated.

Foutbudget is used to balance reliability and development speed. Teams monitor the remaining budget to guide release planning, incident response, and change control. When the budget is exhausted or near exhaustion, organizations tend to slow or halt risky deployments, prioritize reliability work, and, if needed, roll back recent changes.

Measurement relies on service-level indicators (SLIs) and SLOs, together with telemetry and dashboards. The concept originated in Google’s SRE framework and has since been adopted broadly in software engineering and cloud operations as a practical tool for aligning customer expectations with product velocity.

Critiques note that defining appropriate SLOs and budgets can be challenging, and there is a risk of gaming or misinterpretation. Nevertheless, foutbudget remains a widely used principle for embedding reliability into the software development lifecycle.