Multi-Window and Multi-Burn-Rate Alerting
You'll learn how to detect both fast-burning incidents and slow degradations using multiple time windows before your SLOs break.
The Core Problem
A single alerting threshold can't distinguish between a sudden catastrophic failure and a gradual drift toward SLO violation. If you only alert on 5-minute windows, you'll catch fast incidents but miss slow burns. If you only use 24-hour windows, slow problems are visible but fast outages take too long to detect.
How Multi-Window, Multi-Burn-Rate Works
Multi-window means observing your error budget consumption across several time periods simultaneously (e.g., 1 hour, 6 hours, 3 days). Multi-burn-rate means calculating how fast you're burning your error budget relative to your SLO target.
For example, if your monthly error budget allows 0.1% errors:
- Fast burn (1-hour window): If you see 1% errors over 1 hour, you're burning budget 10× faster than sustainable—alert immediately
- Moderate burn (6-hour window): If you see 0.3% errors over 6 hours, you're burning 3× faster—alert with medium priority
- Slow burn (3-day window): If you see 0.15% errors over 3 days, you're burning 1.5× faster—alert as a warning
Why This Matters
This approach gives you proportional response. Critical incidents trigger immediate pages. Gradual degradations create tickets for investigation. You avoid alert fatigue (from noisy short windows) and prevent silent SLO violations (from overly long windows).
Think of it like checking both your speedometer and trip odometer—one tells you if you're currently speeding; the other shows if you'll run out of gas before reaching your destination.
Key Takeaway:
Multi-window, multi-burn-rate alerting detects both sudden outages and gradual SLO erosion by monitoring error budget consumption across multiple time horizons with severity matched to burn rate.