The Nine Nines: Understanding Availability SLAs
You'll learn: What the standard availability tiers (99% to 99.999%) actually mean in real-world downtime allowances.
What Do the "Nines" Mean?
When engineers talk about "three nines" or "five nines" of availability, they're describing the percentage of time a system is operational. Each additional "nine" represents a tenfold reduction in allowed downtime—and typically a significant increase in cost and complexity.
The Standard Tiers
Here's what each availability SLA translates to in actual downtime per year:
- 99% (two nines): ~3.65 days/year (~87.6 hours)
- 99.9% (three nines): ~8.76 hours/year (~43.8 minutes/month)
- 99.99% (four nines): ~52.56 minutes/year (~4.38 minutes/month)
- 99.999% (five nines): ~5.26 minutes/year (~26 seconds/month)
Why This Matters
Think of availability tiers like insurance policies. Two nines might work for an internal tool used during business hours. Three nines suits most customer-facing applications. Four nines is common for critical services like payment processing. Five nines is reserved for emergency systems where downtime could mean life-or-death consequences.
Each additional nine doesn't just cost more—it requires fundamentally different architecture. Moving from 99% to 99.9% might mean adding load balancers. Reaching 99.99% demands multi-region deployments. Achieving 99.999% requires sophisticated automation, redundancy at every layer, and often significant operational overhead.
Choosing Your Target
Your error budget (as we learned earlier) flows directly from your availability SLA. A 99.9% SLA gives you ~43.8 minutes of downtime monthly to "spend" on deployments, experiments, and inevitable failures.
Key Takeaway: Each "nine" in your availability SLA represents roughly 10× less downtime and often 10× more complexity—choose the tier that matches your actual business requirements, not aspirational perfection.