Calculating and Tracking Error Budget
What you'll learn: How to turn your SLO into a concrete error budget number and monitor its consumption throughout the measurement period.
The Core Calculation
An error budget is the mathematical inverse of your SLO—it's the allowed failure you can spend before breaking your reliability promise. The calculation is straightforward:
Error Budget = 100% - SLO Target
Allowed Failures = Total Events × Error Budget
If your SLO is 99.9% availability over 30 days, your error budget is 0.1%. That translates to roughly 43 minutes of downtime per month (0.1% × 43,200 minutes).
Tracking Consumption
Track how much budget you've spent by measuring your actual SLI against the target. Remaining budget at any point:
Budget Remaining = SLO Target - Current SLI Performance
Budget Consumed (%) = (Failures So Far / Allowed Failures) × 100%
Example: Your SLO is 99.95% successful requests over 30 days (error budget = 0.05%). You handle 10 million requests monthly.
- Allowed failures: 10,000,000 × 0.0005 = 5,000 failed requests
- After 15 days, you've had 2,000 failures
- Budget consumed: (2,000 / 5,000) × 100% = 40%
- Status: On track (50% of time passed, 40% of budget used)
The Power of Tracking
Think of this like a monthly data plan—you need visibility into consumption to make smart decisions. Are you burning budget too fast? Freeze risky deployments. Plenty left? Accelerate experimentation.
Key Takeaway: Calculate error budget as allowed failures from your SLO, then continuously track consumption by comparing actual SLI performance to your target—this transforms reliability from philosophy into actionable data.