Service Level Objectives (SLOs): Setting Reliability Targets
What you'll learn: How to define realistic reliability targets that balance user satisfaction with engineering velocity and innovation.
What Are SLOs?
A Service Level Objective (SLO) is a specific target value or range for a Service Level Indicator (SLI). While an SLI measures something (like availability or latency), an SLO sets the goal for that measurement.
Think of it like a thermostat: the thermometer (SLI) tells you the current temperature, but the thermostat setting (SLO) tells you what temperature you're aiming for.
The Balancing Act
SLOs aren't about achieving perfection—they're about finding the sweet spot between:
- Reliability: Keeping your service working well enough to satisfy users
- Development velocity: Having room to innovate, deploy changes, and occasionally fail
Setting an SLO at 100% uptime might sound ideal, but it's unrealistic and counterproductive. It leaves no room for maintenance, deployments, or learning from controlled failures. Most services target something like 99.9% or 99.95% availability—numbers that still allow for brief outages or degraded performance.
Examples of SLOs
- "95% of API requests will complete in under 200ms" (latency SLO)
- "The service will be available 99.9% of the time" (availability SLO)
- "99% of uploads will succeed on the first attempt" (throughput/success rate SLO)
Error Budgets
SLOs create error budgets—the difference between 100% and your SLO is the amount of "unreliability" you can afford. A 99.9% availability SLO means you have a 0.1% error budget to spend on risks, experiments, and rapid deployments.
Key Takeaway: SLOs define concrete reliability targets based on SLI measurements, creating a balance between user expectations and the engineering freedom needed to innovate and improve the system.