SLOs (Service Level Objectives)#
Internal targets for reliability that a service aims to meet.
Important
✨ AI-generated content. This page was written with the assistance of an AI language model and is provided as a learning aid. Despite careful review, it may still contain mistakes, omissions, or out-of-date information. Whether you are new to the topic, a team lead, or a senior practitioner, treat it as a starting point rather than an authoritative reference: read it critically and independently verify anything you act on (code, commands, figures, and factual claims) against official documentation and primary sources before relying on it.
What it is#
A Service Level Objective (SLO) is a measurable target for how a service should perform — usually a percentage or number. It is the goal for reliability, performance or availability that a team commits to.
The SLA, SLO, SLI hierarchy#
The three fit together. The SLA is the external contract with customers; the SLO is the internal target that supports it; and the SLI is the actual measurement used to check performance. The SLO sits in the middle — stricter than the SLA, expressed in terms the SLI can measure.
Examples and error budgets#
Typical SLOs cover availability (“99.9% uptime per month”, SLI = uptime %), latency (“95% of requests under 300 ms”), error rate (“under 0.1% of requests return 5xx in 30 days”) and throughput (“at least 5,000 requests per second at peak”). An SLO of 99.9% implies an error budget of 0.1% allowed downtime — the slack that decides when to prioritise reliability over new features, and that stops teams from the costly, unrealistic chase for 100%.
A worked example#
Suppose an SLA requires 99.5% uptime. The internal SLO is set at 99.9% for margin, and the SLI is measured uptime over the last 30 days. If uptime slips to 99.6%, the SLA is still safe but the SLO is breached — an early warning of reliability risk before any SLA penalty is triggered.
Theme: MLOps, Serving & Monitoring · All terminology
Hint
Mind map — connected ideas
SLI (Service Level Indicator) · SLA (Service Level Agreement) · Model KPIs (Key Performance Indicators) · Monitoring Pipelines · Guardrails (in ML & Data Systems) · Model Stability
Hint
More in MLOps, Serving & Monitoring
AWS SageMaker Endpoints · Caching · Cloud Inference · Cloud Inference with Big Payloads · Compute budgets · Continuous Retraining · Feature Values · Guardrails (in ML & Data Systems) · Inference Cost (Inference $) · Latency Guardrails · Manual review minutes · Model KPIs (Key Performance Indicators) · Model Stability · Monitoring Pipelines
See also
Source article Adapted (context, re-expressed) in our own words from: SLOs (Service Level Objectives) (insightful-data-lab.com).