Error Budget
The acceptable amount of unreliability for a service, calculated as 1 minus the SLO, used to balance reliability and feature development.
Detailed Explanation
The error budget is a concrete, measurable way to balance reliability and velocity. If your SLO is 99.9% availability, your error budget is 0.1%—about 43 minutes of downtime per month. As long as you haven't exhausted the budget, you can ship features aggressively. If the budget is consumed, you must prioritize reliability work.
Error budgets prevent two failure modes: over-investing in reliability (diminishing returns) and under-investing in reliability (user dissatisfaction). They provide a data-driven framework for making engineering decisions about risk.
Why It Matters
Error budgets provide a clear, quantitative framework for balancing feature development with reliability investment.
Real-World Example
A team has a 99.9% SLO (43 minutes downtime/month). After 2 weeks, they have used 30 minutes of their budget (high burn rate). They investigate the cause, fix the reliability issue, and resume feature development.
When to Use
Whenever you have SLOs. Error budgets are the mechanism that makes SLOs actionable and drives engineering decisions.
Advantages
- Quantifies reliability investment
- Balances features and reliability
- Provides clear decision criteria
- Prevents reliability over-investment
- Creates shared understanding across teams
Disadvantages
- Requires accurate SLI measurement
- Budget consumption can be hard to attribute
- Organizational buy-in needed
- Can be gamed if metrics are wrong
- Regular review and adjustment needed
Related Terms
Frequently Asked Questions
How do I calculate an error budget?
Error budget = 1 - SLO. For 99.9% SLO over 30 days: 0.1% × 30 days × 24 hours × 60 minutes = 43.2 minutes. This is the maximum allowed downtime before the budget is exhausted.
What happens when the error budget is exhausted?
The team shifts focus from feature development to reliability work. No new features are shipped until the error budget recovers. This forces reliability investment when it is most needed.
How do I track error budget consumption?
Calculate SLI compliance over the budget window (e.g., 28 days). When SLI falls below the SLO, the error budget is being consumed. Use dashboards (Grafana, Datadog) to visualize budget status.
Can I roll over unused error budget?
Yes. Unused error budget provides a buffer for future incidents. Some teams allow rolling over budget for up to one budget period. This prevents penalizing teams for being reliable.
How do I set the right SLO for my error budget?
Understand user expectations, measure current performance, and set an SLO that is meaningful but achievable. Start with industry standards (99.9% for most web apps) and adjust based on your specific requirements.