Calculating Error Budgets from SLOs
You'll learn: How to translate your SLO percentage into a concrete allowance of "bad minutes" or failed requests you can spend without breaching your reliability target.
The Math Behind Your Error Budget
An error budget is the inverse of your SLO—it's the allowed unreliability. If you have a 99.9% availability SLO, your error budget is 0.1% (100% - 99.9%).
The key is converting that percentage into tangible units over your chosen time window.
Calculation Formulas
For Availability (Time-Based)
Error Budget = (1 - SLO) × Time Period
Example: 99.9% SLO over 30 days:
- Error budget = (1 - 0.999) × 30 days × 24 hours × 60 minutes
- Error budget = 0.001 × 43,200 minutes
- 43.2 minutes of downtime allowed per month
For Request Success (Count-Based)
Allowed Failed Requests = (1 - SLO) × Total Requests
Example: 99.95% success rate with 10 million requests/month:
- Error budget = (1 - 0.9995) × 10,000,000
- 5,000 failed requests allowed per month
Why This Matters
Think of your error budget like a checking account balance. You know exactly how many "failure dollars" you have to spend on incidents, risky deployments, or experiments. Spend wisely—once it's gone, you must freeze changes until the window resets.
These calculations let you answer: "Can we afford this risky deployment?" or "How much outage time remains this quarter?"
Key Takeaway: Error budgets transform abstract SLO percentages into concrete allowances—minutes of downtime or counts of failed requests—that teams can track, spend, and manage strategically.